Improvements by Instructor
Ken Field
Last compiled on 10 February 2026
- Loading Results
- Improvement in learning elements
- Sample Size
- Q10 Figures
- Q10_1
- Q10_2
- Q10_3
- Q10_4
- Final Figure 3A Q10_5
- Q10_6
- Q10_7
- Q10_8
- Final Figure 3B Q10_9
- Q10_10
- Final Figure 3C Q10_11
- Final Figure 3D Q10_12
- Q10_13
- Final Figure 3E Q10_14
- Final Figure 3F Q10_15
- Final Figure 3G Q10_16
- Q10_17
- Final Figure 3H Q10_18
- Q10_19
- Q10_20
- Q10_21
- Q10_22
- Q10_23
- Final Figure 3I Q10_24
- Q10_25
- Summary
- Figures
- Instructor Survey Results
- Presurvey by Semester
- Table S4
IMPORTANT NOTE
This Rmd uses the deidentified results and is safe to share.
Loading Results
Loading in the results without instructor information:
NoDemographicsYear1 <- read_delim("Deidentified Surveys/Year1.NoDemographics.tsv",
delim = "\t", escape_double = FALSE,
trim_ws = TRUE) %>%
rename(Semester = Semester_pre)## Rows: 85 Columns: 151
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (141): ResponseId_pre, Instructor, Semester_pre, Q1_pre, Q8_pre, Q9_1_pr...
## dbl (10): Q19_1_pre, Q19_2_pre, Q19_3_pre, Q19_4_pre, Q19_5_pre, Q19_6_pre,...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.NoDemographicsYear2 <- read_delim("Deidentified Surveys/Year2.NoDemographics.tsv",
delim = "\t", escape_double = FALSE,
trim_ws = TRUE)## Rows: 77 Columns: 151
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (141): ResponseId_pre, Instructor, Semester, Q1_pre, Q8_pre, Q9_1_pre, Q...
## dbl (10): Q19_1_pre, Q19_2_pre, Q19_3_pre, Q19_4_pre, Q19_5_pre, Q19_6_pre,...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.NoDemographicsYear3 <- read_delim("Deidentified Surveys/Year3.NoDemographics.tsv",
delim = "\t", escape_double = FALSE,
trim_ws = TRUE)## Rows: 63 Columns: 151
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (141): ResponseId_pre, Instructor, Semester, Q1_pre, Q8_pre, Q9_1_pre, Q...
## dbl (10): Q19_1_pre, Q19_2_pre, Q19_3_pre, Q19_4_pre, Q19_5_pre, Q19_6_pre,...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.NoDemographics <- bind_rows(NoDemographicsYear1, NoDemographicsYear2, NoDemographicsYear3)
NoDemographicsQuestions <- read_delim("Deidentified Surveys/Year3.NoDemographicsQuestions.tsv",
delim = "\t", escape_double = FALSE,
trim_ws = TRUE)## Rows: 151 Columns: 2
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (2): value, Question
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.Cleaning up the factors and removing anyone who did not agree to the informed consent.
NoDemographics <- NoDemographics %>%
mutate(Semester = factor(Semester, levels = c("Fall 2021", "Spring 2022",
"Fall 2022", "Spring 2023",
"Fall 2023", "Spring 2024"))) %>%
mutate(across(Instructor, str_replace, 'Prof. ', ''))## Warning: There was 1 warning in `mutate()`.
## ℹ In argument: `across(Instructor, str_replace, "Prof. ", "")`.
## Caused by warning:
## ! The `...` argument of `across()` is deprecated as of dplyr 1.1.0.
## Supply arguments directly to `.fns` through an anonymous function instead.
##
## # Previously
## across(a:b, mean, na.rm = TRUE)
##
## # Now
## across(a:b, \(x) mean(x, na.rm = TRUE))NoDemographics <- NoDemographics %>%
filter(Q1_pre == "Agree")We have hypothesized that instructors may be less successful in improving learning elements during the first time that they have taught the class. To test this hypothesis, we will need to code a new variable, that I will call Rookie. To do this, I logged into Cognos and ran a “Class Info by Term and Instructor” report. Note that I had to identify the instructor pseudonyms using the Deidentification Rmd, but I have not included that information here.
- ‘McGonagall’ ‘Fall 2021’
- ‘Dumbledore’ ‘Fall 2021’
- ‘Hagrid’ ‘Spring 2022’
- ‘Lupin’ ‘Fall 2021’
- ‘Sinistra’ ‘Spring 2022’
NoDemographics <- NoDemographics %>%
mutate(Rookie = case_when((Instructor == 'McGonagall') & (Semester == 'Fall 2021') ~ "Rookie",
(Instructor == 'Dumbledore') & (Semester == 'Fall 2021') ~ "Rookie",
(Instructor == 'Hagrid') & (Semester == 'Spring 2022') ~ "Rookie",
(Instructor == 'Lupin') & (Semester == 'Fall 2021') ~ "Rookie",
(Instructor == 'Sinistra') & (Semester == 'Spring 2022') ~ "Rookie",
.default = "Veteran"))Improvement in learning elements
Comparing the responses to Pre10 and Post9:
## [1] "Please look over this inventory of elements that might be included in a course. For each element, give an estimate of your current level of ability before the course begins. Your current level of ability may be a result of courses in high school or college, or it may be a result of other experiences such as jobs or special programs. If students are expected to do the following course elements, what would be their level of expertise? - A scripted lab or project in which the students know the expected outcome"## [1] "A scripted lab or project in which the students know the expected outcome"
## [2] "A lab or project in which only the instructor knows the outcome"
## [3] "A lab or project where no one knows the outcome"
## [4] "At least one project that is assigned and structured by the instructor"
## [5] "A project in which students have some input into the research process and/or what is being studied"
## [6] "A project entirely of student design"
## [7] "Work individually"
## [8] "Work as a whole class"
## [9] "Work in small groups"
## [10] "Become responsible for a part of the project"
## [11] "Read primary scientific literature"
## [12] "Write a research proposal"
## [13] "Collect data"
## [14] "Analyze data"
## [15] "Present results orally"
## [16] "Present results in written papers or reports"
## [17] "Present posters"
## [18] "Critique the work of other students"
## [19] "Listen to lectures"
## [20] "Read a textbook"
## [21] "Work on problem sets"
## [22] "Take tests in class"
## [23] "Discuss reading materials in class"
## [24] "Maintain a lab notebook"
## [25] "Computer modeling"## [1] "Please rate how much learning you gained from each element you experienced in this course. The scale measuring your gain is from (no or very small gain) to (very large gain). Some elements may not have happened at all. If the item is not relevant or you prefer not to answer, please choose the \"\"not applicable\"\" option. If students were expected to do the following course elements, what would be their level of gained experience? - A scripted lab or project in which the students know the expected outcome"## [1] "A scripted lab or project in which the students know the expected outcome"
## [2] "A lab or project in which only the instructor knows the outcome"
## [3] "A lab or project where no one knows the outcome"
## [4] "At least one project that is assigned and structured by the instructor"
## [5] "A project in which students have some input into the research process and/or what is being studied"
## [6] "A project entirely of student design"
## [7] "Work individually"
## [8] "Work as a whole class"
## [9] "Work in small groups"
## [10] "Become responsible for a part of the project"
## [11] "Read primary scientific literature"
## [12] "Write a research proposal"
## [13] "Collect data"
## [14] "Analyze data"
## [15] "Present results orally"
## [16] "Present results in written papers or reports"
## [17] "Present posters"
## [18] "Critique the work of other students"
## [19] "Listen to lectures"
## [20] "Read a textbook"
## [21] "Work on problem sets"
## [22] "Take tests in class"
## [23] "Discuss reading materials in class"
## [24] "Maintain a lab notebook"
## [25] "Computer modeling"## [1] TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE
## [16] TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUENow to compare the pre and post responses for those questions:
## # A tibble: 6 × 53
## Semester Instructor Rookie Q10_1_pre Q10_2_pre Q10_3_pre Q10_4_pre Q10_5_pre
## <fct> <chr> <chr> <chr> <chr> <chr> <chr> <chr>
## 1 Fall 2021 McGonagall Rookie Much Some None Some Extensive
## 2 Fall 2021 McGonagall Rookie Some Much None Extensive Little
## 3 Fall 2021 McGonagall Rookie Some Much None Much Some
## 4 Fall 2021 McGonagall Rookie Little Much Little Some Some
## 5 Fall 2021 McGonagall Rookie Some Some Little Much Much
## 6 Fall 2021 McGonagall Rookie Some Much Little Much Some
## # ℹ 45 more variables: Q10_6_pre <chr>, Q10_7_pre <chr>, Q10_8_pre <chr>,
## # Q10_9_pre <chr>, Q10_10_pre <chr>, Q10_11_pre <chr>, Q10_12_pre <chr>,
## # Q10_13_pre <chr>, Q10_14_pre <chr>, Q10_15_pre <chr>, Q10_16_pre <chr>,
## # Q10_17_pre <chr>, Q10_18_pre <chr>, Q10_19_pre <chr>, Q10_20_pre <chr>,
## # Q10_21_pre <chr>, Q10_22_pre <chr>, Q10_23_pre <chr>, Q10_24_pre <chr>,
## # Q10_25_pre <chr>, Q9_1_post <chr>, Q9_2_post <chr>, Q9_3_post <chr>,
## # Q9_4_post <chr>, Q9_5_post <chr>, Q9_6_post <chr>, Q9_7_post <chr>, …## Warning: `funs()` was deprecated in dplyr 0.8.0.
## ℹ Please use a list of either functions or lambdas:
##
## # Simple named list: list(mean = mean, median = median)
##
## # Auto named with `tibble::lst()`: tibble::lst(mean, median)
##
## # Using lambdas list(~ mean(., trim = .2), ~ median(., na.rm = TRUE))
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.Sample Size
Table S1
Q10Clean %>%
group_by(Instructor, Semester) %>%
count() %>%
print()## # A tibble: 14 × 3
## # Groups: Instructor, Semester [14]
## Instructor Semester n
## <chr> <fct> <int>
## 1 Dumbledore Fall 2021 15
## 2 Dumbledore Fall 2022 24
## 3 Dumbledore Spring 2023 11
## 4 Dumbledore Fall 2023 20
## 5 Hagrid Spring 2022 16
## 6 Hagrid Spring 2024 19
## 7 Lupin Fall 2021 11
## 8 Lupin Fall 2022 13
## 9 McGonagall Fall 2021 8
## 10 McGonagall Spring 2022 25
## 11 Sinistra Spring 2022 10
## 12 Sinistra Spring 2023 28
## 13 Sinistra Fall 2023 10
## 14 Sinistra Spring 2024 14First let’s just look at the contingency tables to see if everything looks right.
## [1] "Rows represents pre-survey response, Columns represent post-survey response."## [1] "First for All sections then for then by Instructor."## [1] "A scripted lab or project in which the students know the expected outcome"##
## None Little Some Much Extensive
## None 1 1 0 0 0
## Little 1 3 8 8 5
## Some 3 7 21 40 14
## Much 3 5 17 33 28
## Extensive 1 4 4 6 2## [1] "By Instructor"## [1] "A scripted lab or project in which the students know the expected outcome"## , , = Dumbledore
##
##
## None Little Some Much Extensive
## None 1 1 0 0 0
## Little 1 0 3 3 3
## Some 1 2 8 7 7
## Much 1 1 3 8 10
## Extensive 0 1 2 1 2
##
## , , = Hagrid
##
##
## None Little Some Much Extensive
## None 0 0 0 0 0
## Little 0 0 2 1 0
## Some 0 1 10 5 0
## Much 1 0 1 5 2
## Extensive 1 2 0 2 0
##
## , , = Lupin
##
##
## None Little Some Much Extensive
## None 0 0 0 0 0
## Little 0 1 1 2 0
## Some 0 1 1 2 2
## Much 0 2 2 8 0
## Extensive 0 0 0 1 0
##
## , , = McGonagall
##
##
## None Little Some Much Extensive
## None 0 0 0 0 0
## Little 0 1 0 1 1
## Some 0 1 1 9 1
## Much 0 0 6 7 4
## Extensive 0 1 0 0 0
##
## , , = Sinistra
##
##
## None Little Some Much Extensive
## None 0 0 0 0 0
## Little 0 1 2 1 1
## Some 2 2 1 17 4
## Much 1 2 5 5 12
## Extensive 0 0 2 2 0Balloon Plot
On these plots, the answers on the x axis are the pre-survey results and on the y-axis are the post survey results. Responses above the “none” level indicate students who felt that they increased in this element.
The pre-survey questions were asked as: Please look over this inventory of elements that might be included in a course. For each element, give an estimate of your current level of ability before the course begins. Your current level of ability may be a result of courses in high school or college, or it may be a result of other experiences such as jobs or special programs. If students are expected to do the following course elements, what would be their level of expertise?
The post-survey questions were asked: Please rate how much learning you gained from each element you experienced in this course. The scale measuring your gain is from (no or very small gain) to (very large gain). Some elements may not have happened at all. If the item is not relevant or you prefer not to answer, please choose the “not applicable” option. If students were expected to do the following course elements, what would be their level of gained experience?
## Dumbledore Hagrid Lupin McGonagall Sinistra
## 70 35 24 33 62Q10 Figures
Q10_1
## [1] "Dumbledore" "Hagrid" "Lupin" "McGonagall" "Sinistra"##
## Call:
## glm(formula = Q9_1_post ~ Q10_1_pre + Instructor * Rookie, data = Q_1Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.752964 0.407954 9.199 <2e-16 ***
## Q10_1_pre 0.124083 0.088566 1.401 0.1627
## InstructorHagrid -0.675089 0.397376 -1.699 0.0909 .
## InstructorLupin -0.324478 0.443863 -0.731 0.4656
## InstructorMcGonagall -0.390722 0.471212 -0.829 0.4080
## InstructorSinistra -0.513693 0.455021 -1.129 0.2602
## RookieVeteran -0.506976 0.324632 -1.562 0.1199
## InstructorHagrid:RookieVeteran 0.127011 0.491822 0.258 0.7965
## InstructorLupin:RookieVeteran -0.005156 0.551577 -0.009 0.9926
## InstructorMcGonagall:RookieVeteran 0.498036 0.536388 0.929 0.3542
## InstructorSinistra:RookieVeteran 0.660594 0.499031 1.324 0.1871
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.099195)
##
## Null deviance: 238.49 on 214 degrees of freedom
## Residual deviance: 224.24 on 204 degrees of freedom
## AIC: 643.19
##
## Number of Fisher Scoring iterations: 2## Start: AIC=643.19
## Q9_1_post ~ Q10_1_pre + Instructor * Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 227.03 637.85
## <none> 224.24 643.19
## - Q10_1_pre 1 226.39 643.25
##
## Step: AIC=637.85
## Q9_1_post ~ Q10_1_pre + Instructor + Rookie
##
## Df Deviance AIC
## <none> 227.03 637.85
## - Instructor 4 235.72 637.92
## - Q10_1_pre 1 229.51 638.18
## - Rookie 1 229.76 638.42##
## Call:
## glm(formula = Q9_1_post ~ Q10_1_pre + Instructor + Rookie, data = Q_1Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.542224 0.333345 10.626 <2e-16 ***
## Q10_1_pre 0.129947 0.086319 1.505 0.1337
## InstructorHagrid -0.545433 0.227678 -2.396 0.0175 *
## InstructorLupin -0.271429 0.256505 -1.058 0.2912
## InstructorMcGonagall -0.003572 0.223378 -0.016 0.9873
## InstructorSinistra 0.035753 0.186927 0.191 0.8485
## RookieVeteran -0.268776 0.170035 -1.581 0.1155
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.091505)
##
## Null deviance: 238.49 on 214 degrees of freedom
## Residual deviance: 227.03 on 208 degrees of freedom
## AIC: 637.85
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## Q9_1_post ~ Q10_1_pre + Instructor + Rookie
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 227.03 637.85
## Q10_1_pre 1 229.51 638.18 2.3299 0.1269
## Instructor 4 235.72 637.92 8.0686 0.0891 .
## Rookie 1 229.76 638.42 2.5674 0.1091
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_2
##
## Call:
## glm(formula = as.numeric(Q9_2_post) ~ as.numeric(Q10_2_pre) +
## Instructor * Rookie, data = Q_2Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.97313 0.36224 10.968 <2e-16 ***
## as.numeric(Q10_2_pre) -0.08628 0.08212 -1.051 0.295
## InstructorHagrid -0.14644 0.37070 -0.395 0.693
## InstructorLupin 0.42369 0.40284 1.052 0.294
## InstructorMcGonagall -0.04614 0.44229 -0.104 0.917
## InstructorSinistra -0.14914 0.42462 -0.351 0.726
## RookieVeteran -0.01550 0.29916 -0.052 0.959
## InstructorHagrid:RookieVeteran -0.51879 0.45848 -1.132 0.259
## InstructorLupin:RookieVeteran -0.68748 0.51255 -1.341 0.181
## InstructorMcGonagall:RookieVeteran 0.00472 0.50451 0.009 0.993
## InstructorSinistra:RookieVeteran 0.17970 0.46754 0.384 0.701
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.9873036)
##
## Null deviance: 212.96 on 214 degrees of freedom
## Residual deviance: 201.41 on 204 degrees of freedom
## AIC: 620.1
##
## Number of Fisher Scoring iterations: 2## Start: AIC=620.1
## as.numeric(Q9_2_post) ~ as.numeric(Q10_2_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 205.35 616.27
## - as.numeric(Q10_2_pre) 1 202.50 619.27
## <none> 201.41 620.10
##
## Step: AIC=616.27
## as.numeric(Q9_2_post) ~ as.numeric(Q10_2_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Instructor 4 211.46 614.57
## - as.numeric(Q10_2_pre) 1 205.95 614.90
## - Rookie 1 206.80 615.78
## <none> 205.35 616.27
##
## Step: AIC=614.57
## as.numeric(Q9_2_post) ~ as.numeric(Q10_2_pre) + Rookie
##
## Df Deviance AIC
## - Rookie 1 212.09 613.21
## - as.numeric(Q10_2_pre) 1 212.36 613.49
## <none> 211.46 614.57
##
## Step: AIC=613.21
## as.numeric(Q9_2_post) ~ as.numeric(Q10_2_pre)
##
## Df Deviance AIC
## - as.numeric(Q10_2_pre) 1 212.96 612.09
## <none> 212.09 613.21
##
## Step: AIC=612.09
## as.numeric(Q9_2_post) ~ 1##
## Call:
## glm(formula = as.numeric(Q9_2_post) ~ 1, data = Q_2Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.61395 0.06803 53.12 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.9951315)
##
## Null deviance: 212.96 on 214 degrees of freedom
## Residual deviance: 212.96 on 214 degrees of freedom
## AIC: 612.09
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_2_post) ~ 1
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 212.96 612.09Q10_3
##
## Call:
## glm(formula = as.numeric(Q9_3_post) ~ as.numeric(Q10_3_pre) +
## Instructor * Rookie, data = Q_3Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.33994 0.28907 15.014 <2e-16 ***
## as.numeric(Q10_3_pre) 0.06789 0.06729 1.009 0.3142
## InstructorHagrid -0.93443 0.36363 -2.570 0.0109 *
## InstructorLupin -0.79165 0.40360 -1.961 0.0511 .
## InstructorMcGonagall -0.31677 0.44213 -0.716 0.4745
## InstructorSinistra -0.10966 0.41385 -0.265 0.7913
## RookieVeteran -0.30624 0.29448 -1.040 0.2996
## InstructorHagrid:RookieVeteran 0.29746 0.45093 0.660 0.5102
## InstructorLupin:RookieVeteran 0.23756 0.51089 0.465 0.6424
## InstructorMcGonagall:RookieVeteran 0.22704 0.50893 0.446 0.6560
## InstructorSinistra:RookieVeteran -0.05254 0.45652 -0.115 0.9085
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.016728)
##
## Null deviance: 231.93 on 221 degrees of freedom
## Residual deviance: 214.53 on 211 degrees of freedom
## AIC: 646.41
##
## Number of Fisher Scoring iterations: 2## Start: AIC=646.41
## as.numeric(Q9_3_post) ~ as.numeric(Q10_3_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 215.37 639.28
## - as.numeric(Q10_3_pre) 1 215.56 645.48
## <none> 214.53 646.41
##
## Step: AIC=639.28
## as.numeric(Q9_3_post) ~ as.numeric(Q10_3_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - as.numeric(Q10_3_pre) 1 216.52 638.46
## - Rookie 1 216.77 638.72
## <none> 215.37 639.28
## - Instructor 4 231.51 647.32
##
## Step: AIC=638.46
## as.numeric(Q9_3_post) ~ Instructor + Rookie
##
## Df Deviance AIC
## - Rookie 1 217.69 637.65
## <none> 216.52 638.46
## - Instructor 4 231.91 645.70
##
## Step: AIC=637.65
## as.numeric(Q9_3_post) ~ Instructor
##
## Df Deviance AIC
## <none> 217.69 637.65
## - Instructor 4 231.93 643.72##
## Call:
## glm(formula = as.numeric(Q9_3_post) ~ Instructor, data = Q_3Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.2429 0.1197 35.442 < 2e-16 ***
## InstructorHagrid -0.6714 0.2073 -3.238 0.00139 **
## InstructorLupin -0.5762 0.2369 -2.432 0.01582 *
## InstructorMcGonagall -0.1179 0.2137 -0.551 0.58190
## InstructorSinistra -0.1445 0.1754 -0.824 0.41103
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.003161)
##
## Null deviance: 231.93 on 221 degrees of freedom
## Residual deviance: 217.69 on 217 degrees of freedom
## AIC: 637.65
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_3_post) ~ Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 217.69 637.65
## Instructor 4 231.93 643.72 14.069 0.007079 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_4
##
## Call:
## glm(formula = Q9_4_post ~ Q10_4_pre + Instructor * Rookie, data = Q_4Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.43056 0.29255 15.145 <2e-16 ***
## Q10_4_pre 0.01083 0.06539 0.166 0.8686
## InstructorHagrid -0.41003 0.27394 -1.497 0.1359
## InstructorLupin -0.01934 0.30318 -0.064 0.9492
## InstructorMcGonagall -0.34618 0.33201 -1.043 0.2983
## InstructorSinistra -0.56847 0.30875 -1.841 0.0670 .
## RookieVeteran -0.28813 0.22105 -1.303 0.1938
## InstructorHagrid:RookieVeteran -0.30318 0.33803 -0.897 0.3708
## InstructorLupin:RookieVeteran -0.16141 0.38320 -0.421 0.6740
## InstructorMcGonagall:RookieVeteran 0.60302 0.37828 1.594 0.1124
## InstructorSinistra:RookieVeteran 0.71255 0.34147 2.087 0.0381 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.5712601)
##
## Null deviance: 137.50 on 223 degrees of freedom
## Residual deviance: 121.68 on 213 degrees of freedom
## AIC: 522.99
##
## Number of Fisher Scoring iterations: 2## Start: AIC=522.99
## Q9_4_post ~ Q10_4_pre + Instructor * Rookie
##
## Df Deviance AIC
## - Q10_4_pre 1 121.69 521.01
## <none> 121.68 522.99
## - Instructor:Rookie 4 128.20 526.68
##
## Step: AIC=521.01
## Q9_4_post ~ Instructor + Rookie + Instructor:Rookie
##
## Df Deviance AIC
## <none> 121.69 521.01
## - Instructor:Rookie 4 128.23 524.73##
## Call:
## glm(formula = Q9_4_post ~ Instructor + Rookie + Instructor:Rookie,
## data = Q_4Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.46667 0.19471 22.940 <2e-16 ***
## InstructorHagrid -0.40417 0.27102 -1.491 0.1374
## InstructorLupin -0.01212 0.29935 -0.040 0.9677
## InstructorMcGonagall -0.34167 0.33014 -1.035 0.3019
## InstructorSinistra -0.56667 0.30786 -1.841 0.0671 .
## RookieVeteran -0.28485 0.21966 -1.297 0.1961
## InstructorHagrid:RookieVeteran -0.30397 0.33723 -0.901 0.3684
## InstructorLupin:RookieVeteran -0.16970 0.37907 -0.448 0.6548
## InstructorMcGonagall:RookieVeteran 0.59985 0.37693 1.591 0.1130
## InstructorSinistra:RookieVeteran 0.71177 0.34066 2.089 0.0379 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.5686639)
##
## Null deviance: 137.50 on 223 degrees of freedom
## Residual deviance: 121.69 on 214 degrees of freedom
## AIC: 521.01
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## Q9_4_post ~ Instructor + Rookie + Instructor:Rookie
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 121.69 521.01
## Instructor:Rookie 4 128.23 524.73 11.72 0.01956 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3A Q10_5
##
## Call:
## glm(formula = as.numeric(Q9_5_post) ~ as.numeric(Q10_5_pre) +
## Instructor * Rookie, data = Q_5Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.44430 0.20738 21.431 < 2e-16 ***
## as.numeric(Q10_5_pre) 0.08339 0.04506 1.850 0.06564 .
## InstructorHagrid -0.69446 0.23574 -2.946 0.00358 **
## InstructorLupin -0.52022 0.26055 -1.997 0.04714 *
## InstructorMcGonagall 0.04511 0.28732 0.157 0.87538
## InstructorSinistra -0.17778 0.26731 -0.665 0.50671
## RookieVeteran -0.04143 0.19133 -0.217 0.82877
## InstructorHagrid:RookieVeteran -0.04631 0.29275 -0.158 0.87446
## InstructorLupin:RookieVeteran -0.32509 0.33156 -0.980 0.32797
## InstructorMcGonagall:RookieVeteran -0.32816 0.32724 -1.003 0.31709
## InstructorSinistra:RookieVeteran 0.16527 0.29601 0.558 0.57720
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4284952)
##
## Null deviance: 112.933 on 222 degrees of freedom
## Residual deviance: 90.841 on 212 degrees of freedom
## AIC: 456.58
##
## Number of Fisher Scoring iterations: 2## Start: AIC=456.58
## as.numeric(Q9_5_post) ~ as.numeric(Q10_5_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 92.149 451.77
## <none> 90.841 456.58
## - as.numeric(Q10_5_pre) 1 92.308 458.15
##
## Step: AIC=451.77
## as.numeric(Q9_5_post) ~ as.numeric(Q10_5_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 92.684 451.06
## <none> 92.149 451.77
## - as.numeric(Q10_5_pre) 1 93.890 453.94
## - Instructor 4 111.434 486.14
##
## Step: AIC=451.06
## as.numeric(Q9_5_post) ~ as.numeric(Q10_5_pre) + Instructor
##
## Df Deviance AIC
## <none> 92.684 451.06
## - as.numeric(Q10_5_pre) 1 94.271 452.85
## - Instructor 4 111.608 484.49##
## Call:
## glm(formula = as.numeric(Q9_5_post) ~ as.numeric(Q10_5_pre) +
## Instructor, data = Q_5Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.40529 0.15219 28.945 < 2e-16 ***
## as.numeric(Q10_5_pre) 0.08558 0.04439 1.928 0.0551 .
## InstructorHagrid -0.71017 0.13589 -5.226 4.06e-07 ***
## InstructorLupin -0.68587 0.15474 -4.433 1.48e-05 ***
## InstructorMcGonagall -0.20312 0.13892 -1.462 0.1452
## InstructorSinistra -0.04189 0.11454 -0.366 0.7149
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4271133)
##
## Null deviance: 112.933 on 222 degrees of freedom
## Residual deviance: 92.684 on 217 degrees of freedom
## AIC: 451.06
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_5_post) ~ as.numeric(Q10_5_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 92.684 451.06
## as.numeric(Q10_5_pre) 1 94.271 452.85 3.788 0.05162 .
## Instructor 4 111.608 484.49 41.434 2.185e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_6
##
## Call:
## glm(formula = as.numeric(Q9_6_post) ~ as.numeric(Q10_6_pre) +
## Instructor * Rookie, data = Q_6Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.59581 0.20125 22.836 < 2e-16 ***
## as.numeric(Q10_6_pre) 0.02952 0.03931 0.751 0.453
## InstructorHagrid -1.22474 0.24752 -4.948 1.54e-06 ***
## InstructorLupin -1.22447 0.27381 -4.472 1.27e-05 ***
## InstructorMcGonagall 0.19800 0.30175 0.656 0.512
## InstructorSinistra 0.22419 0.29032 0.772 0.441
## RookieVeteran -0.13368 0.20103 -0.665 0.507
## InstructorHagrid:RookieVeteran 0.30561 0.30910 0.989 0.324
## InstructorLupin:RookieVeteran 0.13984 0.34636 0.404 0.687
## InstructorMcGonagall:RookieVeteran -0.35225 0.34451 -1.022 0.308
## InstructorSinistra:RookieVeteran -0.07878 0.32002 -0.246 0.806
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4740606)
##
## Null deviance: 155.828 on 220 degrees of freedom
## Residual deviance: 99.553 on 210 degrees of freedom
## AIC: 474.93
##
## Number of Fisher Scoring iterations: 2## Start: AIC=474.93
## as.numeric(Q9_6_post) ~ as.numeric(Q10_6_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 101.264 470.70
## - as.numeric(Q10_6_pre) 1 99.820 473.52
## <none> 99.553 474.93
##
## Step: AIC=470.7
## as.numeric(Q9_6_post) ~ as.numeric(Q10_6_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - as.numeric(Q10_6_pre) 1 101.58 469.39
## - Rookie 1 101.79 469.85
## <none> 101.26 470.70
## - Instructor 4 154.08 555.45
##
## Step: AIC=469.39
## as.numeric(Q9_6_post) ~ Instructor + Rookie
##
## Df Deviance AIC
## - Rookie 1 102.03 468.35
## <none> 101.58 469.39
## - Instructor 4 154.08 553.46
##
## Step: AIC=468.35
## as.numeric(Q9_6_post) ~ Instructor
##
## Df Deviance AIC
## <none> 102.03 468.35
## - Instructor 4 155.83 553.95##
## Call:
## glm(formula = as.numeric(Q9_6_post) ~ Instructor, data = Q_6Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.56522 0.08274 55.177 < 2e-16 ***
## InstructorHagrid -1.02236 0.14262 -7.168 1.19e-11 ***
## InstructorLupin -1.10688 0.16287 -6.796 1.03e-10 ***
## InstructorMcGonagall -0.05007 0.14546 -0.344 0.731
## InstructorSinistra 0.15145 0.12132 1.248 0.213
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4723441)
##
## Null deviance: 155.83 on 220 degrees of freedom
## Residual deviance: 102.03 on 216 degrees of freedom
## AIC: 468.35
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_6_post) ~ Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 102.03 468.35
## Instructor 4 155.83 553.95 93.598 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_7
##
## Call:
## glm(formula = as.numeric(Q9_7_post) ~ as.numeric(Q10_7_pre) +
## Instructor * Rookie, data = Q_7Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.63027 0.35560 10.209 <2e-16 ***
## as.numeric(Q10_7_pre) 0.01092 0.07027 0.155 0.877
## InstructorHagrid -0.29826 0.37484 -0.796 0.427
## InstructorLupin 0.14225 0.41564 0.342 0.733
## InstructorMcGonagall 0.19969 0.45706 0.437 0.663
## InstructorSinistra -0.06958 0.42346 -0.164 0.870
## RookieVeteran 0.14662 0.30349 0.483 0.630
## InstructorHagrid:RookieVeteran -0.36326 0.46468 -0.782 0.435
## InstructorLupin:RookieVeteran -0.65428 0.52326 -1.250 0.213
## InstructorMcGonagall:RookieVeteran -0.45938 0.52000 -0.883 0.378
## InstructorSinistra:RookieVeteran 0.05743 0.46967 0.122 0.903
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.073799)
##
## Null deviance: 238.72 on 222 degrees of freedom
## Residual deviance: 227.65 on 212 degrees of freedom
## AIC: 661.44
##
## Number of Fisher Scoring iterations: 2## Start: AIC=661.44
## as.numeric(Q9_7_post) ~ as.numeric(Q10_7_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 230.48 656.21
## - as.numeric(Q10_7_pre) 1 227.67 659.47
## <none> 227.65 661.44
##
## Step: AIC=656.21
## as.numeric(Q9_7_post) ~ as.numeric(Q10_7_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - as.numeric(Q10_7_pre) 1 230.58 654.30
## - Rookie 1 230.76 654.48
## - Instructor 4 238.69 656.01
## <none> 230.48 656.21
##
## Step: AIC=654.3
## as.numeric(Q9_7_post) ~ Instructor + Rookie
##
## Df Deviance AIC
## - Rookie 1 230.85 652.56
## - Instructor 4 238.69 654.01
## <none> 230.58 654.30
##
## Step: AIC=652.56
## as.numeric(Q9_7_post) ~ Instructor
##
## Df Deviance AIC
## - Instructor 4 238.72 652.03
## <none> 230.85 652.56
##
## Step: AIC=652.03
## as.numeric(Q9_7_post) ~ 1##
## Call:
## glm(formula = as.numeric(Q9_7_post) ~ 1, data = Q_7Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.65022 0.06944 52.57 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.075304)
##
## Null deviance: 238.72 on 222 degrees of freedom
## Residual deviance: 238.72 on 222 degrees of freedom
## AIC: 652.03
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_7_post) ~ 1
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 238.72 652.03Q10_8
##
## Call:
## glm(formula = Q9_8_post ~ Q10_8_pre + Instructor * Rookie, data = Q_8Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.691626 0.345846 10.674 <2e-16 ***
## Q10_8_pre 0.122296 0.069264 1.766 0.0789 .
## InstructorHagrid -0.651588 0.380116 -1.714 0.0880 .
## InstructorLupin -0.726264 0.424558 -1.711 0.0886 .
## InstructorMcGonagall -0.430809 0.467271 -0.922 0.3576
## InstructorSinistra -0.546284 0.431697 -1.265 0.2071
## RookieVeteran -0.384213 0.309969 -1.240 0.2165
## InstructorHagrid:RookieVeteran -0.409305 0.472883 -0.866 0.3877
## InstructorLupin:RookieVeteran 0.072442 0.536343 0.135 0.8927
## InstructorMcGonagall:RookieVeteran 0.003131 0.532158 0.006 0.9953
## InstructorSinistra:RookieVeteran 0.384159 0.478169 0.803 0.4226
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.117372)
##
## Null deviance: 265.98 on 223 degrees of freedom
## Residual deviance: 238.00 on 213 degrees of freedom
## AIC: 673.26
##
## Number of Fisher Scoring iterations: 2## Start: AIC=673.26
## Q9_8_post ~ Q10_8_pre + Instructor * Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 240.72 667.81
## <none> 238.00 673.26
## - Q10_8_pre 1 241.48 674.52
##
## Step: AIC=667.81
## Q9_8_post ~ Q10_8_pre + Instructor + Rookie
##
## Df Deviance AIC
## <none> 240.72 667.81
## - Q10_8_pre 1 244.08 668.92
## - Rookie 1 246.64 671.25
## - Instructor 4 261.44 678.31##
## Call:
## glm(formula = Q9_8_post ~ Q10_8_pre + Instructor + Rookie, data = Q_8Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.70229 0.29263 12.652 < 2e-16 ***
## Q10_8_pre 0.11847 0.06808 1.740 0.08323 .
## InstructorHagrid -0.87314 0.22168 -3.939 0.00011 ***
## InstructorLupin -0.68531 0.25300 -2.709 0.00729 **
## InstructorMcGonagall -0.42748 0.22299 -1.917 0.05655 .
## InstructorSinistra -0.22594 0.18637 -1.212 0.22671
## RookieVeteran -0.38091 0.16491 -2.310 0.02184 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.109311)
##
## Null deviance: 265.98 on 223 degrees of freedom
## Residual deviance: 240.72 on 217 degrees of freedom
## AIC: 667.81
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## Q9_8_post ~ Q10_8_pre + Instructor + Rookie
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 240.72 667.81
## Q10_8_pre 1 244.08 668.92 3.1047 0.0780678 .
## Instructor 4 261.44 678.31 18.4962 0.0009869 ***
## Rookie 1 246.64 671.25 5.4404 0.0196767 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3B Q10_9
##
## Call:
## glm(formula = as.numeric(Q9_9_post) ~ as.numeric(Q10_9_pre) +
## Instructor * Rookie, data = Q_9Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.281521 0.286801 14.929 <2e-16 ***
## as.numeric(Q10_9_pre) 0.080969 0.059918 1.351 0.178
## InstructorHagrid -0.250640 0.228280 -1.098 0.273
## InstructorLupin -0.005838 0.253199 -0.023 0.982
## InstructorMcGonagall -0.125640 0.277765 -0.452 0.651
## InstructorSinistra 0.302699 0.258419 1.171 0.243
## RookieVeteran 0.035382 0.184380 0.192 0.848
## InstructorHagrid:RookieVeteran -0.144028 0.283079 -0.509 0.611
## InstructorLupin:RookieVeteran -0.410401 0.319037 -1.286 0.200
## InstructorMcGonagall:RookieVeteran 0.075144 0.316507 0.237 0.813
## InstructorSinistra:RookieVeteran -0.337224 0.286285 -1.178 0.240
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4006597)
##
## Null deviance: 91.049 on 222 degrees of freedom
## Residual deviance: 84.940 on 212 degrees of freedom
## AIC: 441.6
##
## Number of Fisher Scoring iterations: 2## Start: AIC=441.6
## as.numeric(Q9_9_post) ~ as.numeric(Q10_9_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 86.209 436.91
## - as.numeric(Q10_9_pre) 1 85.671 441.51
## <none> 84.940 441.60
##
## Step: AIC=436.91
## as.numeric(Q9_9_post) ~ as.numeric(Q10_9_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 86.729 436.25
## <none> 86.209 436.91
## - as.numeric(Q10_9_pre) 1 86.987 436.91
## - Instructor 4 90.625 440.05
##
## Step: AIC=436.25
## as.numeric(Q9_9_post) ~ as.numeric(Q10_9_pre) + Instructor
##
## Df Deviance AIC
## <none> 86.729 436.25
## - as.numeric(Q10_9_pre) 1 87.550 436.35
## - Instructor 4 90.667 438.15##
## Call:
## glm(formula = as.numeric(Q9_9_post) ~ as.numeric(Q10_9_pre) +
## Instructor, data = Q_9Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.29193 0.24667 17.399 <2e-16 ***
## as.numeric(Q10_9_pre) 0.08538 0.05955 1.434 0.153
## InstructorHagrid -0.33868 0.13198 -2.566 0.011 *
## InstructorLupin -0.23813 0.15066 -1.581 0.115
## InstructorMcGonagall -0.07063 0.13407 -0.527 0.599
## InstructorSinistra 0.02229 0.11078 0.201 0.841
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.3996713)
##
## Null deviance: 91.049 on 222 degrees of freedom
## Residual deviance: 86.729 on 217 degrees of freedom
## AIC: 436.25
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_9_post) ~ as.numeric(Q10_9_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 86.729 436.25
## as.numeric(Q10_9_pre) 1 87.550 436.35 2.1023 0.14708
## Instructor 4 90.667 438.15 9.9032 0.04209 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_10
##
## Call:
## glm(formula = Q9_10_post ~ Q10_10_pre + Instructor * Rookie,
## data = Q_10Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.33982 0.23833 18.209 <2e-16 ***
## Q10_10_pre 0.11805 0.05380 2.194 0.0293 *
## InstructorHagrid -0.55466 0.22089 -2.511 0.0128 *
## InstructorLupin -0.26658 0.24397 -1.093 0.2758
## InstructorMcGonagall -0.09155 0.27068 -0.338 0.7355
## InstructorSinistra 0.13519 0.24860 0.544 0.5872
## RookieVeteran -0.23338 0.17945 -1.301 0.1948
## InstructorHagrid:RookieVeteran 0.14924 0.27200 0.549 0.5838
## InstructorLupin:RookieVeteran 0.09070 0.30786 0.295 0.7686
## InstructorMcGonagall:RookieVeteran -0.04353 0.30697 -0.142 0.8874
## InstructorSinistra:RookieVeteran -0.02989 0.27496 -0.109 0.9135
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.3695825)
##
## Null deviance: 86.955 on 221 degrees of freedom
## Residual deviance: 77.982 on 211 degrees of freedom
## AIC: 421.75
##
## Number of Fisher Scoring iterations: 2## Start: AIC=421.75
## Q9_10_post ~ Q10_10_pre + Instructor * Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 78.200 414.37
## <none> 77.982 421.75
## - Q10_10_pre 1 79.761 424.76
##
## Step: AIC=414.37
## Q9_10_post ~ Q10_10_pre + Instructor + Rookie
##
## Df Deviance AIC
## <none> 78.200 414.37
## - Rookie 1 79.805 416.88
## - Q10_10_pre 1 80.014 417.46
## - Instructor 4 85.820 427.01##
## Call:
## glm(formula = Q9_10_post ~ Q10_10_pre + Instructor + Rookie,
## data = Q_10Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.31532 0.21321 20.239 < 2e-16 ***
## Q10_10_pre 0.11776 0.05273 2.233 0.026550 *
## InstructorHagrid -0.46562 0.12918 -3.604 0.000389 ***
## InstructorLupin -0.20946 0.14501 -1.444 0.150066
## InstructorMcGonagall -0.12292 0.13053 -0.942 0.347413
## InstructorSinistra 0.10861 0.10616 1.023 0.307423
## RookieVeteran -0.20079 0.09559 -2.101 0.036840 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.3637216)
##
## Null deviance: 86.955 on 221 degrees of freedom
## Residual deviance: 78.200 on 215 degrees of freedom
## AIC: 414.37
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## Q9_10_post ~ Q10_10_pre + Instructor + Rookie
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 78.200 414.37
## Q10_10_pre 1 80.014 417.46 5.0918 0.0240387 *
## Instructor 4 85.820 427.01 20.6415 0.0003729 ***
## Rookie 1 79.805 416.88 4.5101 0.0336950 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3C Q10_11
##
## Call:
## glm(formula = as.numeric(Q9_11_post) ~ as.numeric(Q10_11_pre) +
## Instructor * Rookie, data = Q_11Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.27965 0.21116 20.267 < 2e-16 ***
## as.numeric(Q10_11_pre) 0.12321 0.04795 2.570 0.0109 *
## InstructorHagrid -1.14069 0.24062 -4.741 3.92e-06 ***
## InstructorLupin -0.28565 0.26274 -1.087 0.2782
## InstructorMcGonagall 0.19491 0.29067 0.671 0.5032
## InstructorSinistra -0.24591 0.28054 -0.877 0.3817
## RookieVeteran -0.17009 0.19371 -0.878 0.3809
## InstructorHagrid:RookieVeteran 0.57070 0.29550 1.931 0.0548 .
## InstructorLupin:RookieVeteran 0.20054 0.33289 0.602 0.5475
## InstructorMcGonagall:RookieVeteran -0.29325 0.33030 -0.888 0.3756
## InstructorSinistra:RookieVeteran 0.15717 0.30657 0.513 0.6087
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4357529)
##
## Null deviance: 111.968 on 221 degrees of freedom
## Residual deviance: 91.944 on 211 degrees of freedom
## AIC: 458.32
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q9_11_post) ~ as.numeric(Q10_11_pre) +
## Instructor * Rookie, data = Q_11Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.27965 0.21116 20.267 < 2e-16 ***
## as.numeric(Q10_11_pre) 0.12321 0.04795 2.570 0.0109 *
## InstructorHagrid -1.14069 0.24062 -4.741 3.92e-06 ***
## InstructorLupin -0.28565 0.26274 -1.087 0.2782
## InstructorMcGonagall 0.19491 0.29067 0.671 0.5032
## InstructorSinistra -0.24591 0.28054 -0.877 0.3817
## RookieVeteran -0.17009 0.19371 -0.878 0.3809
## InstructorHagrid:RookieVeteran 0.57070 0.29550 1.931 0.0548 .
## InstructorLupin:RookieVeteran 0.20054 0.33289 0.602 0.5475
## InstructorMcGonagall:RookieVeteran -0.29325 0.33030 -0.888 0.3756
## InstructorSinistra:RookieVeteran 0.15717 0.30657 0.513 0.6087
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4357529)
##
## Null deviance: 111.968 on 221 degrees of freedom
## Residual deviance: 91.944 on 211 degrees of freedom
## AIC: 458.32
##
## Number of Fisher Scoring iterations: 2## Start: AIC=458.32
## as.numeric(Q9_11_post) ~ as.numeric(Q10_11_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 94.947 457.45
## <none> 91.944 458.32
## - as.numeric(Q10_11_pre) 1 94.822 463.16
##
## Step: AIC=457.45
## as.numeric(Q9_11_post) ~ as.numeric(Q10_11_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 94.984 455.54
## <none> 94.947 457.45
## - as.numeric(Q10_11_pre) 1 97.801 462.03
## - Instructor 4 110.364 482.85
##
## Step: AIC=455.54
## as.numeric(Q9_11_post) ~ as.numeric(Q10_11_pre) + Instructor
##
## Df Deviance AIC
## <none> 94.984 455.54
## - as.numeric(Q10_11_pre) 1 97.801 460.03
## - Instructor 4 110.792 481.71##
## Call:
## glm(formula = as.numeric(Q9_11_post) ~ as.numeric(Q10_11_pre) +
## Instructor, data = Q_11Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.15378 0.16087 25.820 < 2e-16 ***
## as.numeric(Q10_11_pre) 0.12074 0.04771 2.531 0.0121 *
## InstructorHagrid -0.78844 0.14130 -5.580 7.17e-08 ***
## InstructorLupin -0.13596 0.15716 -0.865 0.3879
## InstructorMcGonagall -0.02176 0.14234 -0.153 0.8786
## InstructorSinistra -0.12230 0.12016 -1.018 0.3099
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4397417)
##
## Null deviance: 111.968 on 221 degrees of freedom
## Residual deviance: 94.984 on 216 degrees of freedom
## AIC: 455.54
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_11_post) ~ as.numeric(Q10_11_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 94.984 455.54
## as.numeric(Q10_11_pre) 1 97.801 460.03 6.487 0.01087 *
## Instructor 4 110.792 481.71 34.175 6.861e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3D Q10_12
##
## Call:
## glm(formula = as.numeric(Q9_12_post) ~ as.numeric(Q10_12_pre) +
## Instructor * Rookie, data = Q_12Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.68227 0.26299 14.001 <2e-16 ***
## as.numeric(Q10_12_pre) 0.08922 0.05643 1.581 0.1154
## InstructorHagrid -0.41764 0.32819 -1.273 0.2046
## InstructorLupin -0.20645 0.36487 -0.566 0.5721
## InstructorMcGonagall 0.57236 0.40154 1.425 0.1555
## InstructorSinistra 0.63536 0.38642 1.644 0.1016
## RookieVeteran 0.26532 0.26807 0.990 0.3234
## InstructorHagrid:RookieVeteran 0.24092 0.41142 0.586 0.5588
## InstructorLupin:RookieVeteran -0.09403 0.46306 -0.203 0.8393
## InstructorMcGonagall:RookieVeteran -0.76622 0.45755 -1.675 0.0955 .
## InstructorSinistra:RookieVeteran -0.30015 0.42436 -0.707 0.4802
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.8334572)
##
## Null deviance: 200.65 on 220 degrees of freedom
## Residual deviance: 175.03 on 210 degrees of freedom
## AIC: 599.63
##
## Number of Fisher Scoring iterations: 2## Start: AIC=599.63
## as.numeric(Q9_12_post) ~ as.numeric(Q10_12_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 179.04 596.63
## <none> 175.03 599.63
## - as.numeric(Q10_12_pre) 1 177.11 600.24
##
## Step: AIC=596.63
## as.numeric(Q9_12_post) ~ as.numeric(Q10_12_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 179.65 595.39
## <none> 179.04 596.63
## - as.numeric(Q10_12_pre) 1 182.04 598.31
## - Instructor 4 192.65 604.83
##
## Step: AIC=595.39
## as.numeric(Q9_12_post) ~ as.numeric(Q10_12_pre) + Instructor
##
## Df Deviance AIC
## <none> 179.65 595.39
## - as.numeric(Q10_12_pre) 1 183.07 597.56
## - Instructor 4 195.63 606.22##
## Call:
## glm(formula = as.numeric(Q9_12_post) ~ as.numeric(Q10_12_pre) +
## Instructor, data = Q_12Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.83756 0.17173 22.347 <2e-16 ***
## as.numeric(Q10_12_pre) 0.11071 0.05472 2.023 0.0443 *
## InstructorHagrid -0.35398 0.19049 -1.858 0.0645 .
## InstructorLupin -0.32267 0.21711 -1.486 0.1387
## InstructorMcGonagall -0.02164 0.19493 -0.111 0.9117
## InstructorSinistra 0.38944 0.16298 2.389 0.0177 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.8355852)
##
## Null deviance: 200.65 on 220 degrees of freedom
## Residual deviance: 179.65 on 215 degrees of freedom
## AIC: 595.39
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_12_post) ~ as.numeric(Q10_12_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 179.65 595.39
## as.numeric(Q10_12_pre) 1 183.07 597.56 4.1689 0.0411729 *
## Instructor 4 195.63 606.22 18.8297 0.0008489 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_13
##
## Call:
## glm(formula = as.numeric(Q9_13_post) ~ as.numeric(Q10_13_pre) +
## Instructor * Rookie, data = Q_13Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.46594 0.25719 17.364 < 2e-16 ***
## as.numeric(Q10_13_pre) 0.10545 0.05128 2.056 0.040969 *
## InstructorHagrid -0.81526 0.23498 -3.469 0.000632 ***
## InstructorLupin -0.35189 0.25850 -1.361 0.174878
## InstructorMcGonagall -0.03912 0.28555 -0.137 0.891160
## InstructorSinistra -0.06311 0.27440 -0.230 0.818317
## RookieVeteran -0.13648 0.18965 -0.720 0.472522
## InstructorHagrid:RookieVeteran 0.08619 0.29279 0.294 0.768773
## InstructorLupin:RookieVeteran -0.42763 0.32730 -1.307 0.192781
## InstructorMcGonagall:RookieVeteran -0.10575 0.32593 -0.324 0.745910
## InstructorSinistra:RookieVeteran -0.03076 0.30314 -0.101 0.919263
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.422651)
##
## Null deviance: 111.279 on 221 degrees of freedom
## Residual deviance: 89.179 on 211 degrees of freedom
## AIC: 451.54
##
## Number of Fisher Scoring iterations: 2## Start: AIC=451.54
## as.numeric(Q9_13_post) ~ as.numeric(Q10_13_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 90.208 446.09
## <none> 89.179 451.54
## - as.numeric(Q10_13_pre) 1 90.967 453.94
##
## Step: AIC=446.09
## as.numeric(Q9_13_post) ~ as.numeric(Q10_13_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## <none> 90.208 446.09
## - Rookie 1 91.848 448.09
## - as.numeric(Q10_13_pre) 1 92.510 449.68
## - Instructor 4 108.278 478.62##
## Call:
## glm(formula = as.numeric(Q9_13_post) ~ as.numeric(Q10_13_pre) +
## Instructor + Rookie, data = Q_13Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.47608 0.21579 20.743 < 2e-16 ***
## as.numeric(Q10_13_pre) 0.11686 0.04989 2.342 0.020076 *
## InstructorHagrid -0.78363 0.13645 -5.743 3.15e-08 ***
## InstructorLupin -0.60128 0.15537 -3.870 0.000144 ***
## InstructorMcGonagall -0.12205 0.13687 -0.892 0.373537
## InstructorSinistra -0.08739 0.11464 -0.762 0.446688
## RookieVeteran -0.20228 0.10231 -1.977 0.049301 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4195731)
##
## Null deviance: 111.279 on 221 degrees of freedom
## Residual deviance: 90.208 on 215 degrees of freedom
## AIC: 446.09
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_13_post) ~ as.numeric(Q10_13_pre) + Instructor +
## Rookie
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 90.208 446.09
## as.numeric(Q10_13_pre) 1 92.510 449.68 5.594 0.01802 *
## Instructor 4 108.278 478.62 40.534 3.356e-08 ***
## Rookie 1 91.848 448.09 4.000 0.04550 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3E Q10_14
##
## Call:
## glm(formula = as.numeric(Q9_14_post) ~ as.numeric(Q10_14_pre) +
## Instructor * Rookie, data = Q_14Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.06464 0.26191 15.519 < 2e-16 ***
## as.numeric(Q10_14_pre) 0.17294 0.05170 3.345 0.000973 ***
## InstructorHagrid -0.76250 0.23730 -3.213 0.001518 **
## InstructorLupin -0.59030 0.26044 -2.267 0.024432 *
## InstructorMcGonagall -0.17463 0.28763 -0.607 0.544424
## InstructorSinistra 0.09825 0.27682 0.355 0.723011
## RookieVeteran -0.03582 0.19135 -0.187 0.851707
## InstructorHagrid:RookieVeteran 0.09653 0.29485 0.327 0.743697
## InstructorLupin:RookieVeteran -0.21761 0.32968 -0.660 0.509926
## InstructorMcGonagall:RookieVeteran 0.05014 0.32841 0.153 0.878794
## InstructorSinistra:RookieVeteran -0.11486 0.30531 -0.376 0.707149
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4295947)
##
## Null deviance: 117.279 on 221 degrees of freedom
## Residual deviance: 90.644 on 211 degrees of freedom
## AIC: 455.16
##
## Number of Fisher Scoring iterations: 2## Start: AIC=455.16
## as.numeric(Q9_14_post) ~ as.numeric(Q10_14_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 91.096 448.26
## <none> 90.644 455.16
## - as.numeric(Q10_14_pre) 1 95.452 464.63
##
## Step: AIC=448.26
## as.numeric(Q9_14_post) ~ as.numeric(Q10_14_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 91.246 446.63
## <none> 91.096 448.26
## - as.numeric(Q10_14_pre) 1 96.218 458.40
## - Instructor 4 110.604 483.34
##
## Step: AIC=446.63
## as.numeric(Q9_14_post) ~ as.numeric(Q10_14_pre) + Instructor
##
## Df Deviance AIC
## <none> 91.246 446.63
## - as.numeric(Q10_14_pre) 1 96.494 457.04
## - Instructor 4 111.281 482.69##
## Call:
## glm(formula = as.numeric(Q9_14_post) ~ as.numeric(Q10_14_pre) +
## Instructor, data = Q_14Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.02124 0.20017 20.089 < 2e-16 ***
## as.numeric(Q10_14_pre) 0.17710 0.05025 3.524 0.000518 ***
## InstructorHagrid -0.70074 0.13478 -5.199 4.64e-07 ***
## InstructorLupin -0.70010 0.15396 -4.547 9.05e-06 ***
## InstructorMcGonagall -0.13586 0.13727 -0.990 0.323411
## InstructorSinistra -0.00201 0.11442 -0.018 0.986001
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.4224366)
##
## Null deviance: 117.279 on 221 degrees of freedom
## Residual deviance: 91.246 on 216 degrees of freedom
## AIC: 446.63
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_14_post) ~ as.numeric(Q10_14_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 91.246 446.63
## as.numeric(Q10_14_pre) 1 96.494 457.04 12.413 0.0004263 ***
## Instructor 4 111.281 482.69 44.065 6.218e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3F Q10_15
##
## Call:
## glm(formula = as.numeric(Q9_15_post) ~ as.numeric(Q10_15_pre) +
## Instructor * Rookie, data = Q_15Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.74613 0.26933 13.909 <2e-16 ***
## as.numeric(Q10_15_pre) 0.16210 0.05521 2.936 0.0037 **
## InstructorHagrid -0.22229 0.30708 -0.724 0.4699
## InstructorLupin -0.77538 0.34155 -2.270 0.0242 *
## InstructorMcGonagall -0.70873 0.37671 -1.881 0.0613 .
## InstructorSinistra 0.33516 0.34941 0.959 0.3385
## RookieVeteran 0.04236 0.24959 0.170 0.8654
## InstructorHagrid:RookieVeteran -0.03143 0.38570 -0.081 0.9351
## InstructorLupin:RookieVeteran -0.10020 0.43016 -0.233 0.8160
## InstructorMcGonagall:RookieVeteran 0.03308 0.42930 0.077 0.9387
## InstructorSinistra:RookieVeteran -0.20139 0.38622 -0.521 0.6026
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7295826)
##
## Null deviance: 185.70 on 220 degrees of freedom
## Residual deviance: 153.21 on 210 degrees of freedom
## AIC: 570.21
##
## Number of Fisher Scoring iterations: 2## Start: AIC=570.21
## as.numeric(Q9_15_post) ~ as.numeric(Q10_15_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 153.49 562.61
## <none> 153.21 570.21
## - as.numeric(Q10_15_pre) 1 159.50 577.10
##
## Step: AIC=562.61
## as.numeric(Q9_15_post) ~ as.numeric(Q10_15_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 153.50 560.63
## <none> 153.49 562.61
## - as.numeric(Q10_15_pre) 1 159.88 569.63
## - Instructor 4 180.98 591.02
##
## Step: AIC=560.63
## as.numeric(Q9_15_post) ~ as.numeric(Q10_15_pre) + Instructor
##
## Df Deviance AIC
## <none> 153.50 560.63
## - as.numeric(Q10_15_pre) 1 159.89 567.64
## - Instructor 4 181.72 589.93##
## Call:
## glm(formula = as.numeric(Q9_15_post) ~ as.numeric(Q10_15_pre) +
## Instructor, data = Q_15Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.77979 0.19029 19.863 < 2e-16 ***
## as.numeric(Q10_15_pre) 0.16192 0.05413 2.991 0.003101 **
## InstructorHagrid -0.24960 0.17776 -1.404 0.161717
## InstructorLupin -0.83978 0.20197 -4.158 4.64e-05 ***
## InstructorMcGonagall -0.68521 0.18271 -3.750 0.000227 ***
## InstructorSinistra 0.16869 0.14897 1.132 0.258751
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7139563)
##
## Null deviance: 185.7 on 220 degrees of freedom
## Residual deviance: 153.5 on 215 degrees of freedom
## AIC: 560.63
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_15_post) ~ as.numeric(Q10_15_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 153.50 560.63
## as.numeric(Q10_15_pre) 1 159.89 567.64 9.012 0.002682 **
## Instructor 4 181.72 589.93 37.301 1.562e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3G Q10_16
##
## Call:
## glm(formula = as.numeric(Q9_16_post) ~ as.numeric(Q10_16_pre) +
## Instructor * Rookie, data = Q_16Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.05737 0.26301 15.426 < 2e-16 ***
## as.numeric(Q10_16_pre) 0.12279 0.05185 2.368 0.01878 *
## InstructorHagrid -0.73824 0.27650 -2.670 0.00817 **
## InstructorLupin -0.53897 0.30490 -1.768 0.07855 .
## InstructorMcGonagall 0.16857 0.33618 0.501 0.61660
## InstructorSinistra -0.25029 0.31354 -0.798 0.42560
## RookieVeteran -0.16582 0.22411 -0.740 0.46018
## InstructorHagrid:RookieVeteran 0.35367 0.34381 1.029 0.30480
## InstructorLupin:RookieVeteran -0.06642 0.38632 -0.172 0.86366
## InstructorMcGonagall:RookieVeteran -0.27760 0.38418 -0.723 0.47075
## InstructorSinistra:RookieVeteran 0.42594 0.34717 1.227 0.22121
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.5895735)
##
## Null deviance: 145.09 on 222 degrees of freedom
## Residual deviance: 124.99 on 212 degrees of freedom
## AIC: 527.74
##
## Number of Fisher Scoring iterations: 2## Start: AIC=527.74
## as.numeric(Q9_16_post) ~ as.numeric(Q10_16_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 127.58 524.32
## <none> 124.99 527.74
## - as.numeric(Q10_16_pre) 1 128.30 531.57
##
## Step: AIC=524.32
## as.numeric(Q9_16_post) ~ as.numeric(Q10_16_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 127.70 522.53
## <none> 127.58 524.32
## - as.numeric(Q10_16_pre) 1 130.87 528.00
## - Instructor 4 140.80 538.30
##
## Step: AIC=522.53
## as.numeric(Q9_16_post) ~ as.numeric(Q10_16_pre) + Instructor
##
## Df Deviance AIC
## <none> 127.70 522.53
## - as.numeric(Q10_16_pre) 1 130.96 526.15
## - Instructor 4 141.19 536.92##
## Call:
## glm(formula = as.numeric(Q9_16_post) ~ as.numeric(Q10_16_pre) +
## Instructor, data = Q_16Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.93094 0.19434 20.228 < 2e-16 ***
## as.numeric(Q10_16_pre) 0.12178 0.05175 2.353 0.01950 *
## InstructorHagrid -0.50671 0.15959 -3.175 0.00172 **
## InstructorLupin -0.53505 0.18181 -2.943 0.00361 **
## InstructorMcGonagall -0.03752 0.16239 -0.231 0.81748
## InstructorSinistra 0.09755 0.13433 0.726 0.46850
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.5884753)
##
## Null deviance: 145.09 on 222 degrees of freedom
## Residual deviance: 127.70 on 217 degrees of freedom
## AIC: 522.53
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_16_post) ~ as.numeric(Q10_16_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 127.70 522.53
## as.numeric(Q10_16_pre) 1 130.96 526.15 5.6198 0.0177590 *
## Instructor 4 141.19 536.92 22.3935 0.0001673 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_17
##
## Call:
## glm(formula = as.numeric(Q9_17_post) ~ as.numeric(Q10_17_pre) +
## Instructor * Rookie, data = Q_17Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.15172 0.28469 14.583 <2e-16 ***
## as.numeric(Q10_17_pre) 0.03448 0.05478 0.629 0.5297
## InstructorHagrid -0.55043 0.30744 -1.790 0.0748 .
## InstructorLupin -0.53417 0.33591 -1.590 0.1133
## InstructorMcGonagall -0.13879 0.37038 -0.375 0.7082
## InstructorSinistra 0.44483 0.34584 1.286 0.1998
## RookieVeteran -0.15486 0.24645 -0.628 0.5304
## InstructorHagrid:RookieVeteran 0.43378 0.38267 1.134 0.2583
## InstructorLupin:RookieVeteran -0.03829 0.42538 -0.090 0.9284
## InstructorMcGonagall:RookieVeteran 0.12181 0.42435 0.287 0.7744
## InstructorSinistra:RookieVeteran -0.18440 0.38233 -0.482 0.6301
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7156195)
##
## Null deviance: 168.20 on 222 degrees of freedom
## Residual deviance: 151.71 on 212 degrees of freedom
## AIC: 570.95
##
## Number of Fisher Scoring iterations: 2## Start: AIC=570.95
## as.numeric(Q9_17_post) ~ as.numeric(Q10_17_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 153.52 565.59
## - as.numeric(Q10_17_pre) 1 152.00 569.37
## <none> 151.71 570.95
##
## Step: AIC=565.59
## as.numeric(Q9_17_post) ~ as.numeric(Q10_17_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 153.85 564.07
## - as.numeric(Q10_17_pre) 1 154.03 564.33
## <none> 153.52 565.59
## - Instructor 4 167.66 577.24
##
## Step: AIC=564.07
## as.numeric(Q9_17_post) ~ as.numeric(Q10_17_pre) + Instructor
##
## Df Deviance AIC
## - as.numeric(Q10_17_pre) 1 154.30 562.72
## <none> 153.85 564.07
## - Instructor 4 167.79 575.41
##
## Step: AIC=562.72
## as.numeric(Q9_17_post) ~ Instructor
##
## Df Deviance AIC
## <none> 154.3 562.72
## - Instructor 4 168.2 573.95##
## Call:
## glm(formula = as.numeric(Q9_17_post) ~ Instructor, data = Q_17Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.14286 0.10055 41.200 <2e-16 ***
## InstructorHagrid -0.28571 0.17417 -1.640 0.1023
## InstructorLupin -0.51786 0.19900 -2.602 0.0099 **
## InstructorMcGonagall -0.04911 0.17953 -0.274 0.7847
## InstructorSinistra 0.27650 0.14672 1.885 0.0608 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7077875)
##
## Null deviance: 168.2 on 222 degrees of freedom
## Residual deviance: 154.3 on 218 degrees of freedom
## AIC: 562.72
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_17_post) ~ Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 154.3 562.72
## Instructor 4 168.2 573.95 19.235 0.0007067 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3H Q10_18
##
## Call:
## glm(formula = as.numeric(Q9_18_post) ~ as.numeric(Q10_18_pre) +
## Instructor * Rookie, data = Q_18Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.03117 0.25434 15.850 < 2e-16 ***
## as.numeric(Q10_18_pre) 0.14559 0.05493 2.651 0.008649 **
## InstructorHagrid -1.15275 0.30117 -3.828 0.000171 ***
## InstructorLupin -1.15551 0.32748 -3.528 0.000513 ***
## InstructorMcGonagall 0.33665 0.36102 0.933 0.352142
## InstructorSinistra -0.49706 0.33860 -1.468 0.143610
## RookieVeteran -0.13822 0.24128 -0.573 0.567342
## InstructorHagrid:RookieVeteran 0.32535 0.37523 0.867 0.386894
## InstructorLupin:RookieVeteran 0.39785 0.41508 0.958 0.338915
## InstructorMcGonagall:RookieVeteran -0.58714 0.41366 -1.419 0.157272
## InstructorSinistra:RookieVeteran 0.35685 0.37355 0.955 0.340521
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.6798709)
##
## Null deviance: 180.78 on 220 degrees of freedom
## Residual deviance: 142.77 on 210 degrees of freedom
## AIC: 554.61
##
## Number of Fisher Scoring iterations: 2## Start: AIC=554.61
## as.numeric(Q9_18_post) ~ as.numeric(Q10_18_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 147.24 553.42
## <none> 142.77 554.61
## - as.numeric(Q10_18_pre) 1 147.55 559.89
##
## Step: AIC=553.42
## as.numeric(Q9_18_post) ~ as.numeric(Q10_18_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 147.26 551.46
## <none> 147.24 553.42
## - as.numeric(Q10_18_pre) 1 151.65 557.95
## - Instructor 4 176.32 585.25
##
## Step: AIC=551.46
## as.numeric(Q9_18_post) ~ as.numeric(Q10_18_pre) + Instructor
##
## Df Deviance AIC
## <none> 147.26 551.46
## - as.numeric(Q10_18_pre) 1 151.66 555.96
## - Instructor 4 177.67 584.94##
## Call:
## glm(formula = as.numeric(Q9_18_post) ~ as.numeric(Q10_18_pre) +
## Instructor, data = Q_18Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.94638 0.17471 22.588 < 2e-16 ***
## as.numeric(Q10_18_pre) 0.13692 0.05403 2.534 0.012 *
## InstructorHagrid -0.93611 0.17592 -5.321 2.58e-07 ***
## InstructorLupin -0.90625 0.19658 -4.610 6.90e-06 ***
## InstructorMcGonagall -0.10102 0.17744 -0.569 0.570
## InstructorSinistra -0.20248 0.14694 -1.378 0.170
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.6849462)
##
## Null deviance: 180.78 on 220 degrees of freedom
## Residual deviance: 147.26 on 215 degrees of freedom
## AIC: 551.46
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_18_post) ~ as.numeric(Q10_18_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 147.26 551.46
## as.numeric(Q10_18_pre) 1 151.66 555.96 6.506 0.01075 *
## Instructor 4 177.67 584.94 41.483 2.135e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_19
##
## Call:
## glm(formula = as.numeric(Q9_19_post) ~ as.numeric(Q10_19_pre) +
## Instructor * Rookie, data = Q_19Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.85246 0.41216 6.921 5.58e-11 ***
## as.numeric(Q10_19_pre) 0.05784 0.07378 0.784 0.434
## InstructorHagrid 0.28033 0.41607 0.674 0.501
## InstructorLupin 0.49318 0.45757 1.078 0.282
## InstructorMcGonagall 0.09209 0.52561 0.175 0.861
## InstructorSinistra 0.73353 0.46957 1.562 0.120
## RookieVeteran 0.02591 0.34016 0.076 0.939
## InstructorHagrid:RookieVeteran -0.38685 0.51734 -0.748 0.455
## InstructorLupin:RookieVeteran -0.36769 0.57720 -0.637 0.525
## InstructorMcGonagall:RookieVeteran -0.11738 0.59829 -0.196 0.845
## InstructorSinistra:RookieVeteran -0.08196 0.52014 -0.158 0.875
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.286019)
##
## Null deviance: 285.77 on 216 degrees of freedom
## Residual deviance: 264.92 on 206 degrees of freedom
## AIC: 683.12
##
## Number of Fisher Scoring iterations: 2## Start: AIC=683.12
## as.numeric(Q9_19_post) ~ as.numeric(Q10_19_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 265.95 675.96
## - as.numeric(Q10_19_pre) 1 265.71 681.76
## <none> 264.92 683.12
##
## Step: AIC=675.96
## as.numeric(Q9_19_post) ~ as.numeric(Q10_19_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - as.numeric(Q10_19_pre) 1 266.78 674.63
## - Rookie 1 266.80 674.65
## <none> 265.95 675.96
## - Instructor 4 284.19 682.36
##
## Step: AIC=674.63
## as.numeric(Q9_19_post) ~ Instructor + Rookie
##
## Df Deviance AIC
## - Rookie 1 267.52 673.24
## <none> 266.78 674.63
## - Instructor 4 285.56 681.39
##
## Step: AIC=673.24
## as.numeric(Q9_19_post) ~ Instructor
##
## Df Deviance AIC
## <none> 267.52 673.24
## - Instructor 4 285.77 679.56##
## Call:
## glm(formula = as.numeric(Q9_19_post) ~ Instructor, data = Q_19Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.08824 0.13623 22.670 < 2e-16 ***
## InstructorHagrid 0.08824 0.23595 0.374 0.708809
## InstructorLupin 0.28676 0.26671 1.075 0.283519
## InstructorMcGonagall 0.01176 0.24621 0.048 0.961934
## InstructorSinistra 0.68226 0.19810 3.444 0.000691 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.261904)
##
## Null deviance: 285.77 on 216 degrees of freedom
## Residual deviance: 267.52 on 212 degrees of freedom
## AIC: 673.24
##
## Number of Fisher Scoring iterations: 2Q10_20
##
## Call:
## glm(formula = as.numeric(Q9_20_post) ~ as.numeric(Q10_20_pre) +
## Instructor * Rookie, data = Q_20Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.053198 0.475795 4.315 2.61e-05 ***
## as.numeric(Q10_20_pre) 0.007534 0.079992 0.094 0.925
## InstructorHagrid 0.354167 0.465834 0.760 0.448
## InstructorLupin -0.457392 0.556868 -0.821 0.413
## InstructorMcGonagall -0.080320 0.650097 -0.124 0.902
## InstructorSinistra 0.169492 0.557586 0.304 0.761
## RookieVeteran -0.044034 0.390323 -0.113 0.910
## InstructorHagrid:RookieVeteran -0.247380 0.593023 -0.417 0.677
## InstructorLupin:RookieVeteran 0.495595 0.673021 0.736 0.462
## InstructorMcGonagall:RookieVeteran 0.091397 0.725305 0.126 0.900
## InstructorSinistra:RookieVeteran 0.414770 0.610076 0.680 0.497
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.488009)
##
## Null deviance: 286.34 on 193 degrees of freedom
## Residual deviance: 272.31 on 183 degrees of freedom
## AIC: 640.33
##
## Number of Fisher Scoring iterations: 2## Start: AIC=640.33
## as.numeric(Q9_20_post) ~ as.numeric(Q10_20_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 274.73 634.05
## - as.numeric(Q10_20_pre) 1 272.32 638.34
## <none> 272.31 640.33
##
## Step: AIC=634.05
## as.numeric(Q9_20_post) ~ as.numeric(Q10_20_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - as.numeric(Q10_20_pre) 1 274.79 632.09
## - Rookie 1 274.90 632.17
## - Instructor 4 286.01 633.85
## <none> 274.73 634.05
##
## Step: AIC=632.09
## as.numeric(Q9_20_post) ~ Instructor + Rookie
##
## Df Deviance AIC
## - Rookie 1 274.95 630.20
## - Instructor 4 286.03 631.86
## <none> 274.79 632.09
##
## Step: AIC=630.2
## as.numeric(Q9_20_post) ~ Instructor
##
## Df Deviance AIC
## - Instructor 4 286.33 630.07
## <none> 274.95 630.20
##
## Step: AIC=630.07
## as.numeric(Q9_20_post) ~ 1##
## Call:
## glm(formula = as.numeric(Q9_20_post) ~ 1, data = Q_20Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.21134 0.08745 25.29 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.483601)
##
## Null deviance: 286.34 on 193 degrees of freedom
## Residual deviance: 286.34 on 193 degrees of freedom
## AIC: 630.07
##
## Number of Fisher Scoring iterations: 2Q10_21
##
## Call:
## glm(formula = as.numeric(Q9_21_post) ~ as.numeric(Q10_21_pre) +
## Instructor * Rookie, data = Q_21Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.42955 0.56458 4.303 2.79e-05 ***
## as.numeric(Q10_21_pre) -0.04584 0.10647 -0.431 0.667
## InstructorHagrid 0.25709 0.51895 0.495 0.621
## InstructorLupin -0.59826 0.60500 -0.989 0.324
## InstructorMcGonagall -0.23701 0.70246 -0.337 0.736
## InstructorSinistra 0.96944 0.60597 1.600 0.111
## RookieVeteran -0.07718 0.42416 -0.182 0.856
## InstructorHagrid:RookieVeteran 0.13551 0.65434 0.207 0.836
## InstructorLupin:RookieVeteran 0.57253 0.73102 0.783 0.435
## InstructorMcGonagall:RookieVeteran 0.06343 0.78388 0.081 0.936
## InstructorSinistra:RookieVeteran -0.42177 0.66698 -0.632 0.528
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.738379)
##
## Null deviance: 328.99 on 186 degrees of freedom
## Residual deviance: 305.95 on 176 degrees of freedom
## AIC: 646.75
##
## Number of Fisher Scoring iterations: 2## Start: AIC=646.75
## as.numeric(Q9_21_post) ~ as.numeric(Q10_21_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 308.77 640.46
## - as.numeric(Q10_21_pre) 1 306.28 644.95
## <none> 305.95 646.75
##
## Step: AIC=640.46
## as.numeric(Q9_21_post) ~ as.numeric(Q10_21_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 308.83 638.50
## - as.numeric(Q10_21_pre) 1 309.59 638.96
## <none> 308.77 640.46
## - Instructor 4 327.76 643.62
##
## Step: AIC=638.5
## as.numeric(Q9_21_post) ~ as.numeric(Q10_21_pre) + Instructor
##
## Df Deviance AIC
## - as.numeric(Q10_21_pre) 1 309.63 636.98
## <none> 308.83 638.50
## - Instructor 4 327.77 641.63
##
## Step: AIC=636.98
## as.numeric(Q9_21_post) ~ Instructor
##
## Df Deviance AIC
## <none> 309.63 636.98
## - Instructor 4 328.99 640.33##
## Call:
## glm(formula = as.numeric(Q9_21_post) ~ Instructor, data = Q_21Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.1935 0.1657 13.242 <2e-16 ***
## InstructorHagrid 0.3422 0.2970 1.152 0.2508
## InstructorLupin -0.2412 0.3293 -0.732 0.4649
## InstructorMcGonagall -0.1935 0.3090 -0.626 0.5319
## InstructorSinistra 0.6104 0.2466 2.475 0.0142 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.701282)
##
## Null deviance: 328.99 on 186 degrees of freedom
## Residual deviance: 309.63 on 182 degrees of freedom
## AIC: 636.98
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_21_post) ~ Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 309.63 636.98
## Instructor 4 328.99 640.33 11.342 0.02298 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_22
##
## Call:
## glm(formula = as.numeric(Q9_22_post) ~ as.numeric(Q10_22_pre) +
## Instructor * Rookie, data = Q_22Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.24233 0.48358 4.637 6.8e-06 ***
## as.numeric(Q10_22_pre) -0.04336 0.09085 -0.477 0.634
## InstructorHagrid 0.26445 0.47129 0.561 0.575
## InstructorLupin -0.68304 0.55680 -1.227 0.222
## InstructorMcGonagall -0.06021 0.64820 -0.093 0.926
## InstructorSinistra 0.15402 0.55491 0.278 0.782
## RookieVeteran -0.18312 0.39141 -0.468 0.640
## InstructorHagrid:RookieVeteran 0.19454 0.59955 0.324 0.746
## InstructorLupin:RookieVeteran 0.57652 0.67164 0.858 0.392
## InstructorMcGonagall:RookieVeteran 0.12879 0.72293 0.178 0.859
## InstructorSinistra:RookieVeteran 0.51159 0.60882 0.840 0.402
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.47466)
##
## Null deviance: 283.45 on 189 degrees of freedom
## Residual deviance: 263.96 on 179 degrees of freedom
## AIC: 625.67
##
## Number of Fisher Scoring iterations: 2## Start: AIC=625.67
## as.numeric(Q9_22_post) ~ as.numeric(Q10_22_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 265.64 618.87
## - as.numeric(Q10_22_pre) 1 264.30 623.91
## <none> 263.96 625.67
##
## Step: AIC=618.87
## as.numeric(Q9_22_post) ~ as.numeric(Q10_22_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 265.80 616.98
## - as.numeric(Q10_22_pre) 1 265.95 617.09
## <none> 265.64 618.87
## - Instructor 4 282.61 622.63
##
## Step: AIC=616.98
## as.numeric(Q9_22_post) ~ as.numeric(Q10_22_pre) + Instructor
##
## Df Deviance AIC
## - as.numeric(Q10_22_pre) 1 266.07 615.17
## <none> 265.80 616.98
## - Instructor 4 282.85 620.80
##
## Step: AIC=615.17
## as.numeric(Q9_22_post) ~ Instructor
##
## Df Deviance AIC
## <none> 266.07 615.17
## - Instructor 4 283.45 619.20##
## Call:
## glm(formula = as.numeric(Q9_22_post) ~ Instructor, data = Q_22Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.92063 0.15109 12.712 <2e-16 ***
## InstructorHagrid 0.42419 0.26911 1.576 0.1167
## InstructorLupin -0.30159 0.30218 -0.998 0.3196
## InstructorMcGonagall 0.03937 0.28347 0.139 0.8897
## InstructorSinistra 0.57937 0.22469 2.578 0.0107 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.438202)
##
## Null deviance: 283.45 on 189 degrees of freedom
## Residual deviance: 266.07 on 185 degrees of freedom
## AIC: 615.17
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_22_post) ~ Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 266.07 615.17
## Instructor 4 283.45 619.20 12.026 0.01716 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_23
##
## Call:
## glm(formula = Q9_23_post ~ Q10_23_pre + Instructor * Rookie,
## data = Q_23Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.84991 0.34111 11.286 < 2e-16 ***
## Q10_23_pre 0.16520 0.06814 2.425 0.0162 *
## InstructorHagrid -1.40637 0.31746 -4.430 1.51e-05 ***
## InstructorLupin -0.37396 0.35133 -1.064 0.2884
## InstructorMcGonagall -0.51072 0.38578 -1.324 0.1870
## InstructorSinistra -0.81072 0.35980 -2.253 0.0253 *
## RookieVeteran -0.44716 0.25680 -1.741 0.0831 .
## InstructorHagrid:RookieVeteran -0.18521 0.39797 -0.465 0.6421
## InstructorLupin:RookieVeteran 0.25889 0.44459 0.582 0.5610
## InstructorMcGonagall:RookieVeteran 0.53050 0.44167 1.201 0.2310
## InstructorSinistra:RookieVeteran 0.90419 0.39794 2.272 0.0241 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7747479)
##
## Null deviance: 227.41 on 221 degrees of freedom
## Residual deviance: 163.47 on 211 degrees of freedom
## AIC: 586.07
##
## Number of Fisher Scoring iterations: 2## Start: AIC=586.07
## Q9_23_post ~ Q10_23_pre + Instructor * Rookie
##
## Df Deviance AIC
## <none> 163.47 586.07
## - Instructor:Rookie 4 169.87 586.59
## - Q10_23_pre 1 168.03 590.17##
## Call:
## glm(formula = Q9_23_post ~ Q10_23_pre + Instructor * Rookie,
## data = Q_23Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.84991 0.34111 11.286 < 2e-16 ***
## Q10_23_pre 0.16520 0.06814 2.425 0.0162 *
## InstructorHagrid -1.40637 0.31746 -4.430 1.51e-05 ***
## InstructorLupin -0.37396 0.35133 -1.064 0.2884
## InstructorMcGonagall -0.51072 0.38578 -1.324 0.1870
## InstructorSinistra -0.81072 0.35980 -2.253 0.0253 *
## RookieVeteran -0.44716 0.25680 -1.741 0.0831 .
## InstructorHagrid:RookieVeteran -0.18521 0.39797 -0.465 0.6421
## InstructorLupin:RookieVeteran 0.25889 0.44459 0.582 0.5610
## InstructorMcGonagall:RookieVeteran 0.53050 0.44167 1.201 0.2310
## InstructorSinistra:RookieVeteran 0.90419 0.39794 2.272 0.0241 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7747479)
##
## Null deviance: 227.41 on 221 degrees of freedom
## Residual deviance: 163.47 on 211 degrees of freedom
## AIC: 586.07
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## Q9_23_post ~ Q10_23_pre + Instructor * Rookie
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 163.47 586.07
## Q10_23_pre 1 168.03 590.17 6.1004 0.01352 *
## Instructor:Rookie 4 169.87 586.59 8.5171 0.07437 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Final Figure 3I Q10_24
##
## Call:
## glm(formula = as.numeric(Q9_24_post) ~ as.numeric(Q10_24_pre) +
## Instructor * Rookie, data = Q_24Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.196899 0.274343 15.298 < 2e-16 ***
## as.numeric(Q10_24_pre) 0.135510 0.049514 2.737 0.00675 **
## InstructorHagrid -2.022321 0.309163 -6.541 4.73e-10 ***
## InstructorLupin -0.004022 0.329162 -0.012 0.99026
## InstructorMcGonagall 0.129633 0.363266 0.357 0.72157
## InstructorSinistra -0.176327 0.339986 -0.519 0.60458
## RookieVeteran -0.014813 0.241749 -0.061 0.95120
## InstructorHagrid:RookieVeteran -0.221958 0.390883 -0.568 0.57076
## InstructorLupin:RookieVeteran -0.216870 0.417956 -0.519 0.60440
## InstructorMcGonagall:RookieVeteran -0.466106 0.416191 -1.120 0.26405
## InstructorSinistra:RookieVeteran -0.056323 0.377196 -0.149 0.88145
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.6870052)
##
## Null deviance: 254.48 on 216 degrees of freedom
## Residual deviance: 141.52 on 206 degrees of freedom
## AIC: 547.07
##
## Number of Fisher Scoring iterations: 2## Start: AIC=547.07
## as.numeric(Q9_24_post) ~ as.numeric(Q10_24_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 142.52 540.59
## <none> 141.52 547.07
## - as.numeric(Q10_24_pre) 1 146.67 552.82
##
## Step: AIC=540.59
## as.numeric(Q9_24_post) ~ as.numeric(Q10_24_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## - Rookie 1 143.67 540.33
## <none> 142.52 540.59
## - as.numeric(Q10_24_pre) 1 147.57 546.15
## - Instructor 4 249.56 654.16
##
## Step: AIC=540.33
## as.numeric(Q9_24_post) ~ as.numeric(Q10_24_pre) + Instructor
##
## Df Deviance AIC
## <none> 143.67 540.33
## - as.numeric(Q10_24_pre) 1 148.27 545.18
## - Instructor 4 250.19 652.70##
## Call:
## glm(formula = as.numeric(Q9_24_post) ~ as.numeric(Q10_24_pre) +
## Instructor, data = Q_24Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.21601 0.18679 22.571 < 2e-16 ***
## as.numeric(Q10_24_pre) 0.12611 0.04849 2.601 0.00996 **
## InstructorHagrid -2.13638 0.18009 -11.863 < 2e-16 ***
## InstructorLupin -0.11573 0.19550 -0.592 0.55451
## InstructorMcGonagall -0.21962 0.17609 -1.247 0.21370
## InstructorSinistra -0.22379 0.14454 -1.548 0.12307
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.6808933)
##
## Null deviance: 254.48 on 216 degrees of freedom
## Residual deviance: 143.67 on 211 degrees of freedom
## AIC: 540.33
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_24_post) ~ as.numeric(Q10_24_pre) + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 143.67 540.33
## as.numeric(Q10_24_pre) 1 148.27 545.18 6.847 0.008878 **
## Instructor 4 250.19 652.70 120.368 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Q10_25
##
## Call:
## glm(formula = as.numeric(Q9_25_post) ~ as.numeric(Q10_25_pre) +
## Instructor * Rookie, data = Q_25Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.80658 0.31626 12.036 < 2e-16 ***
## as.numeric(Q10_25_pre) 0.19682 0.06801 2.894 0.00424 **
## InstructorHagrid -1.08951 0.39011 -2.793 0.00575 **
## InstructorLupin -1.19846 0.45384 -2.641 0.00894 **
## InstructorMcGonagall -1.53356 0.67405 -2.275 0.02398 *
## InstructorSinistra -0.47483 0.46956 -1.011 0.31317
## RookieVeteran -0.47209 0.31744 -1.487 0.13859
## InstructorHagrid:RookieVeteran -0.47186 0.49139 -0.960 0.33811
## InstructorLupin:RookieVeteran 0.85494 0.56061 1.525 0.12888
## InstructorMcGonagall:RookieVeteran 0.80611 0.72355 1.114 0.26661
## InstructorSinistra:RookieVeteran 0.50965 0.51422 0.991 0.32285
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.122431)
##
## Null deviance: 281.38 on 205 degrees of freedom
## Residual deviance: 218.87 on 195 degrees of freedom
## AIC: 621.09
##
## Number of Fisher Scoring iterations: 2## Start: AIC=621.09
## as.numeric(Q9_25_post) ~ as.numeric(Q10_25_pre) + Instructor *
## Rookie
##
## Df Deviance AIC
## - Instructor:Rookie 4 226.66 620.29
## <none> 218.87 621.09
## - as.numeric(Q10_25_pre) 1 228.27 627.75
##
## Step: AIC=620.29
## as.numeric(Q9_25_post) ~ as.numeric(Q10_25_pre) + Instructor +
## Rookie
##
## Df Deviance AIC
## <none> 226.66 620.29
## - Rookie 1 229.28 620.66
## - as.numeric(Q10_25_pre) 1 233.43 624.35
## - Instructor 4 274.22 651.53##
## Call:
## glm(formula = as.numeric(Q9_25_post) ~ as.numeric(Q10_25_pre) +
## Instructor + Rookie, data = Q_25Clean)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.72720 0.23574 15.810 < 2e-16 ***
## as.numeric(Q10_25_pre) 0.16325 0.06696 2.438 0.015646 *
## InstructorHagrid -1.28376 0.23174 -5.540 9.52e-08 ***
## InstructorLupin -0.65351 0.26402 -2.475 0.014152 *
## InstructorMcGonagall -0.83873 0.24615 -3.407 0.000794 ***
## InstructorSinistra -0.05260 0.19240 -0.273 0.784856
## RookieVeteran -0.27850 0.18345 -1.518 0.130569
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.138992)
##
## Null deviance: 281.38 on 205 degrees of freedom
## Residual deviance: 226.66 on 199 degrees of freedom
## AIC: 620.29
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Q9_25_post) ~ as.numeric(Q10_25_pre) + Instructor +
## Rookie
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 226.66 620.29
## as.numeric(Q10_25_pre) 1 233.43 624.35 6.063 0.0138 *
## Instructor 4 274.22 651.53 39.239 6.219e-08 ***
## Rookie 1 229.28 620.66 2.372 0.1235
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Summary
For each question, generalized linear model selection using AIC determined if the student’s post-survey response was dependent on (1) that student’s pre-survey response, (2) which instructor they had, and (3) whether it was that instructor’s first semester teaching in the CURE Lab (Rookie status).
Removing “Q10_02”, “Q10_19”, “Q10_20”, “Q10_21”, “Q10_22”, “Q10_25” *See below (not emphasized by any instructors)
The glm was improved by including pre-survey response, instructor, rookie status, and the interaction for the following question:
23
- [23] “Discuss reading materials in class”
The glm was improved by including instructor, rookie status, and the interaction, but not pre-survey response, for the following question: 4
- [4] “At least one project that is assigned and structured by the instructor”
The glm was improved by including pre-survey response, instructor, and rookie status (but not the interaction) for the following questions:
1, 8, 10, 13, 25
- [1] “A scripted lab or project in which the students know the
expected outcome”
- [8] “Work as a whole class”
- [10] “Become responsible for a part of the project”
- [13] “Collect data”
- [25] “Computer modeling” *Removed
The following questions showed dependence on pre-survey response and instructor, but not rookie status:
5, 9, 11, 12, 14, 15, 16, 18, 24
- [5] “A project in which students have some input into the research process and/or what is being studied”
- [9] “Work in small groups”
- [11] “Read primary scientific literature”
- [12] “Write a research proposal”
- [14] “Analyze data”
- [15] “Present results orally”
- [16] “Present results in written papers or reports”
- [18] “Critique the work of other students”
- [24] “Maintain a lab notebook”
The following questions showed dependence on instructor, but not pre-survey response or rookie status:
3, 6, 17, 19, 21, 22
- [3] “A lab or project where no one knows the outcome”
- [6] “A project entirely of student design”
- [17] “Present posters”
- [19] “Listen to lectures” *Removed
- [21] “Work on problem sets” *Removed
- [22] “Take tests in class” *Removed
And the responses to these survey questions were independent of pre-survey responses, the instructor, and the rookie status:
2, 7, 20
- [2] “A lab or project in which only the instructor knows the
outcome” *Removed
- [7] “Work individually”
- [20] “Read a textbook” *Removed
It is interesting to note that for every question that was significantly impacted by rookie status, the veteran status had a negative estimate. This means that students in an instructor’s first semester perceived large gains than in subsequent semesters. This does not support our hypothesis that experience would make the instructors more proficient at assisting student improvement in the learning elements.
Figures
For questions 4 and 23 we want to show the interaction between instructor and rookie. Although for 4, the presurvey response wasn’t important, so we will use a cat plot.
## Warning: Q10_23_pre and Instructor and Rookie are not included in an interaction
## with one another in the model.
## Warning: Q10_23_pre and Instructor and Rookie are not included in an interaction
## with one another in the model.## Warning: Q10_1_pre and Instructor and Rookie are not included in an interaction with
## one another in the model.
## Warning: Q10_1_pre and Instructor and Rookie are not included in an interaction with
## one another in the model.## Warning: Q10_8_pre and Instructor and Rookie are not included in an interaction with
## one another in the model.
## Warning: Q10_8_pre and Instructor and Rookie are not included in an interaction with
## one another in the model.## Warning: Q10_10_pre and Instructor and Rookie are not included in an interaction
## with one another in the model.
## Warning: Q10_10_pre and Instructor and Rookie are not included in an interaction
## with one another in the model.## Using data Q_13Clean from global environment. This could cause incorrect
## results if Q_13Clean has been altered since the model was fit. You can
## manually provide the data to the "data =" argument.## Warning: Q10_13_pre and Instructor and Rookie are not included in an interaction
## with one another in the model.
## Using data Q_13Clean from global environment. This could cause incorrect
## results if Q_13Clean has been altered since the model was fit. You can
## manually provide the data to the "data =" argument.## Warning: Q10_13_pre and Instructor and Rookie are not included in an interaction
## with one another in the model.Removed Figure Q10_2
This was not emphasized
Removed Figure Q10_21
This was not emphasized.
For questions where the instructor was included in the final model, but not Rookie status or the pre-survey response (3, 6, and 17) and ALSO the question where it did not depend on any of the variables (7):
Final Figure 2
## [1] "A lab or project where no one knows the outcome"
## [2] "A project entirely of student design"
## [3] "Present posters"
## [4] "Work individually"Note that we lost “Figure 4” because it is now incorporated into Figure 2. I am instead going to convert Figure 3B (with 9 parts!) into Figure 4.
Final Figure 4
Figure 4 is the combined Cat and Interact plots.
## [1] "A scripted lab or project in which the students know the expected outcome"
## [2] "Work as a whole class"
## [3] "Become responsible for a part of the project"
## [4] "Collect data"
## [5] "Discuss reading materials in class"
## [6] "At least one project that is assigned and structured by the instructor"
Figure 4A
Plotted separately.
## [1] "A scripted lab or project in which the students know the expected outcome"
## [2] "Work as a whole class"
## [3] "Become responsible for a part of the project"
## [4] "Collect data"
## [5] "Discuss reading materials in class"
## [6] "At least one project that is assigned and structured by the instructor"
Instructor Survey Results
We will now examine how the responses to the CURE Instructor Survey might correlate with the student perceptions of gain corresponding to each course element. We predict that instructors who placed a greater emphasis on a course element will demonstrate greater perceived gains due to that element.
The five instructors filled out the forms in the summer of 2022. Note that it is possible that instructor emphasis shifted slightly in later years, but none of those shifts were dramatic enough to affect these results. The forms were PDFs that were sent by email, so I will manually enter the results here.
First to visualize how much variability there is in each question.
I want to see if certain questions and instructors cluster together to help decide if there are questions that are more interesting to look at than others.
Final Figure 5A
The following six questions have at least three instructors who selected NA/None for the Emphasis:
## # A tibble: 6 × 1
## Question
## <chr>
## 1 A lab or project in which only the instructor knows the outcome
## 2 Listen to lectures
## 3 Read a textbook
## 4 Work on problem sets
## 5 Take tests in class
## 6 Computer modelingIt is also interesting to note that the following four questions had Major Emphasis in all five sections:
## # A tibble: 4 × 1
## Question
## <chr>
## 1 A lab or project where no one knows the outcome
## 2 Read primary scientific literature
## 3 Collect data
## 4 Analyze dataNow to test whether student-perceived gains are correlated with instructor emphasis on each course element.
Looking at all the questions together, including those in the previous list where we did not emphasize them.
## tibble [5,433 × 4] (S3: tbl_df/tbl/data.frame)
## $ Instructor: chr [1:5433] "McGonagall" "McGonagall" "McGonagall" "McGonagall" ...
## $ Question : chr [1:5433] "Q10_01" "Q10_02" "Q10_03" "Q10_04" ...
## $ Gain : Ord.factor w/ 5 levels "None"<"Little"<..: 4 4 3 4 5 5 3 2 4 5 ...
## $ Emphasis : num [1:5433] 1 1 3 0 3 3 1 2 3 3 ...## Instructor Question Gain Emphasis
## Length:5433 Length:5433 None : 304 Min. :0.000
## Class :character Class :character Little : 446 1st Qu.:1.000
## Mode :character Mode :character Some : 893 Median :2.000
## Much :1773 Mean :1.896
## Extensive:2017 3rd Qu.:3.000
## Max. :3.000##
## Call:
## glm(formula = as.numeric(Gain) ~ Emphasis, data = Q10_post_merged)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.99042 0.02630 113.71 <2e-16 ***
## Emphasis 0.46651 0.01178 39.61 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.049255)
##
## Null deviance: 7344.9 on 5432 degrees of freedom
## Residual deviance: 5698.5 on 5431 degrees of freedom
## AIC: 15683
##
## Number of Fisher Scoring iterations: 2## Single term deletions
##
## Model:
## as.numeric(Gain) ~ Emphasis
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 5698.5 15683
## Emphasis 1 7344.9 17060 1378.9 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Self-reported gain in these skills is highly dependent on the emphasis placed by the instructor (p < 2e-16).
Now to explore whether this also depended on the instructor.
Final Figure 5B
## Start: AIC=15438.6
## as.numeric(Gain) ~ Emphasis * Instructor
##
## Df Deviance AIC
## - Emphasis:Instructor 4 5439.3 15438
## <none> 5431.4 15439
##
## Step: AIC=15438.46
## as.numeric(Gain) ~ Emphasis + Instructor
##
## Df Deviance AIC
## <none> 5439.3 15438
## - Instructor 4 5698.5 15683
## - Emphasis 1 7052.2 16847##
## Call:
## glm(formula = as.numeric(Gain) ~ Emphasis + Instructor, data = Q10_post_merged)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.153854 0.031792 99.204 < 2e-16 ***
## Emphasis 0.463762 0.011561 40.116 < 2e-16 ***
## InstructorHagrid -0.537189 0.042100 -12.760 < 2e-16 ***
## InstructorLupin -0.461789 0.047890 -9.643 < 2e-16 ***
## InstructorMcGonagall -0.183935 0.043178 -4.260 2.08e-05 ***
## InstructorSinistra 0.006216 0.035478 0.175 0.861
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.002262)
##
## Null deviance: 7344.9 on 5432 degrees of freedom
## Residual deviance: 5439.3 on 5427 degrees of freedom
## AIC: 15438
##
## Number of Fisher Scoring iterations: 2## Using data Q10_post_merged from global environment. This could cause
## incorrect results if Q10_post_merged has been altered since the model was
## fit. You can manually provide the data to the "data =" argument.## Warning: Emphasis and Instructor are not included in an interaction with one another
## in the model.
## Single term deletions
##
## Model:
## Gain ~ Emphasis + Instructor
## Df Deviance AIC scaled dev. Pr(>Chi)
## <none> 5379.9 15383
## Emphasis 3 7052.2 16847 1470.51 < 2.2e-16 ***
## Instructor 4 5676.9 15667 291.93 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Warning: Emphasis and Instructor are not included in an interaction with one another
## in the model.
Just to show the interaction:
interact_plot(model = interact_model, pred = Emphasis, modx = Instructor,
interval = TRUE, x.label = "Instructor Emphasis",
y.label = "Student Percieved Gain")## Using data Q10_post_merged from global environment. This could cause
## incorrect results if Q10_post_merged has been altered since the model was
## fit. You can manually provide the data to the "data =" argument.
Presurvey by Semester
Responses on the presurvey had an effect on the student perceived gain for several questions. We want to check to see if the presurvey responses were significantly different in the two semesters, because that could be a confounding factor.
##
## Call:
## glm(formula = as.numeric(Q10_1_pre) ~ FallSpring, data = Q_1Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.29897 0.08399 39.278 <2e-16 ***
## FallSpringSpring 0.22645 0.11337 1.997 0.0471 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.6842893)
##
## Null deviance: 148.48 on 214 degrees of freedom
## Residual deviance: 145.75 on 213 degrees of freedom
## AIC: 532.57
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_2_pre) ~ FallSpring, data = Q_2Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.15152 0.08484 37.146 <2e-16 ***
## FallSpringSpring 0.21917 0.11551 1.898 0.0591 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7126179)
##
## Null deviance: 154.35 on 214 degrees of freedom
## Residual deviance: 151.79 on 213 degrees of freedom
## AIC: 541.29
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_3_pre) ~ FallSpring, data = Q_3Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.0792 0.1020 20.376 < 2e-16 ***
## FallSpringSpring 0.4662 0.1382 3.373 0.000878 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.051665)
##
## Null deviance: 243.33 on 221 degrees of freedom
## Residual deviance: 231.37 on 220 degrees of freedom
## AIC: 645.18
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_4_pre) ~ FallSpring, data = Q_4Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.62376 0.07891 45.922 <2e-16 ***
## FallSpringSpring 0.17299 0.10649 1.624 0.106
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.6289264)
##
## Null deviance: 141.28 on 223 degrees of freedom
## Residual deviance: 139.62 on 222 degrees of freedom
## AIC: 535.8
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_5_pre) ~ FallSpring, data = Q_5Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.88119 0.09865 29.205 <2e-16 ***
## FallSpringSpring 0.30734 0.13338 2.304 0.0221 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.9829782)
##
## Null deviance: 222.46 on 222 degrees of freedom
## Residual deviance: 217.24 on 221 degrees of freedom
## AIC: 633.01
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_6_pre) ~ FallSpring, data = Q_6Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.5300 0.1203 21.037 <2e-16 ***
## FallSpringSpring 0.2634 0.1625 1.621 0.107
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.446323)
##
## Null deviance: 320.54 on 220 degrees of freedom
## Residual deviance: 316.74 on 219 degrees of freedom
## AIC: 712.72
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_7_pre) ~ FallSpring, data = Q_7Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.75248 0.10078 37.236 <2e-16 ***
## FallSpringSpring 0.04261 0.13625 0.313 0.755
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.025742)
##
## Null deviance: 226.79 on 222 degrees of freedom
## Residual deviance: 226.69 on 221 degrees of freedom
## AIC: 642.51
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_8_pre) ~ FallSpring, data = Q_8Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.5545 0.1055 33.692 <2e-16 ***
## FallSpringSpring -0.2780 0.1424 -1.953 0.0521 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.124109)
##
## Null deviance: 253.84 on 223 degrees of freedom
## Residual deviance: 249.55 on 222 degrees of freedom
## AIC: 665.88
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_9_pre) ~ FallSpring, data = Q_9Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.03960 0.07204 56.072 <2e-16 ***
## FallSpringSpring 0.05056 0.09740 0.519 0.604
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.5242072)
##
## Null deviance: 115.99 on 222 degrees of freedom
## Residual deviance: 115.85 on 221 degrees of freedom
## AIC: 492.81
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_10_pre) ~ FallSpring, data = Q_10Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.82178 0.07838 48.762 <2e-16 ***
## FallSpringSpring 0.22780 0.10616 2.146 0.033 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.6204298)
##
## Null deviance: 139.35 on 221 degrees of freedom
## Residual deviance: 136.49 on 220 degrees of freedom
## AIC: 528.03
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_11_pre) ~ FallSpring, data = Q_11Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.89000 0.09173 31.506 < 2e-16 ***
## FallSpringSpring 0.70836 0.12374 5.725 3.37e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.8414076)
##
## Null deviance: 212.68 on 221 degrees of freedom
## Residual deviance: 185.11 on 220 degrees of freedom
## AIC: 595.66
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_12_pre) ~ FallSpring, data = Q_12Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.4100 0.1128 21.367 <2e-16 ***
## FallSpringSpring 0.3751 0.1524 2.461 0.0146 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.272161)
##
## Null deviance: 286.31 on 220 degrees of freedom
## Residual deviance: 278.60 on 219 degrees of freedom
## AIC: 684.36
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_13_pre) ~ FallSpring, data = Q_13Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.73267 0.08774 42.541 <2e-16 ***
## FallSpringSpring -0.03019 0.11885 -0.254 0.8
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7775974)
##
## Null deviance: 171.12 on 221 degrees of freedom
## Residual deviance: 171.07 on 220 degrees of freedom
## AIC: 578.16
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_14_pre) ~ FallSpring, data = Q_14Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.71287 0.08724 42.560 <2e-16 ***
## FallSpringSpring -0.03519 0.11816 -0.298 0.766
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7686501)
##
## Null deviance: 169.17 on 221 degrees of freedom
## Residual deviance: 169.10 on 220 degrees of freedom
## AIC: 575.59
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_15_pre) ~ FallSpring, data = Q_15Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.1600 0.1070 29.533 <2e-16 ***
## FallSpringSpring 0.1375 0.1446 0.951 0.343
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.144882)
##
## Null deviance: 251.76 on 220 degrees of freedom
## Residual deviance: 250.73 on 219 degrees of freedom
## AIC: 661.06
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_16_pre) ~ FallSpring, data = Q_16Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.32000 0.09988 33.241 <2e-16 ***
## FallSpringSpring -0.13301 0.13448 -0.989 0.324
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.9975529)
##
## Null deviance: 221.43 on 222 degrees of freedom
## Residual deviance: 220.46 on 221 degrees of freedom
## AIC: 636.29
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_17_pre) ~ FallSpring, data = Q_17Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.3069 0.1048 31.542 <2e-16 ***
## FallSpringSpring -0.2741 0.1417 -1.934 0.0544 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.110199)
##
## Null deviance: 249.51 on 222 degrees of freedom
## Residual deviance: 245.35 on 221 degrees of freedom
## AIC: 660.15
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_18_pre) ~ FallSpring, data = Q_18Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.6263 0.1028 25.542 < 2e-16 ***
## FallSpringSpring 0.5295 0.1384 3.826 0.00017 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.046633)
##
## Null deviance: 244.53 on 220 degrees of freedom
## Residual deviance: 229.21 on 219 degrees of freedom
## AIC: 641.23
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_19_pre) ~ FallSpring, data = Q_19Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.6837 0.1065 34.583 <2e-16 ***
## FallSpringSpring 0.3499 0.1438 2.433 0.0158 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.111904)
##
## Null deviance: 245.64 on 216 degrees of freedom
## Residual deviance: 239.06 on 215 degrees of freedom
## AIC: 642.83
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_20_pre) ~ FallSpring, data = Q_20Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.835165 0.116069 33.042 <2e-16 ***
## FallSpringSpring -0.009922 0.159293 -0.062 0.95
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.225947)
##
## Null deviance: 235.39 on 193 degrees of freedom
## Residual deviance: 235.38 on 192 degrees of freedom
## AIC: 594.06
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_21_pre) ~ FallSpring, data = Q_21Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.87500 0.09964 38.89 <2e-16 ***
## FallSpringSpring -0.01641 0.13695 -0.12 0.905
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.8737578)
##
## Null deviance: 161.66 on 186 degrees of freedom
## Residual deviance: 161.65 on 185 degrees of freedom
## AIC: 509.44
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_22_pre) ~ FallSpring, data = Q_22Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.1236 0.1052 39.186 <2e-16 ***
## FallSpringSpring -0.1830 0.1443 -1.268 0.206
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.9855533)
##
## Null deviance: 186.87 on 189 degrees of freedom
## Residual deviance: 185.28 on 188 degrees of freedom
## AIC: 540.42
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_23_pre) ~ FallSpring, data = Q_23Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.95050 0.08739 45.208 <2e-16 ***
## FallSpringSpring 0.02471 0.11837 0.209 0.835
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.7712641)
##
## Null deviance: 169.71 on 221 degrees of freedom
## Residual deviance: 169.68 on 220 degrees of freedom
## AIC: 576.34
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_24_pre) ~ FallSpring, data = Q_24Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.32673 0.11569 28.755 <2e-16 ***
## FallSpringSpring -0.02501 0.15824 -0.158 0.875
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.351895)
##
## Null deviance: 290.69 on 216 degrees of freedom
## Residual deviance: 290.66 on 215 degrees of freedom
## AIC: 685.24
##
## Number of Fisher Scoring iterations: 2##
## Call:
## glm(formula = as.numeric(Q10_25_pre) ~ FallSpring, data = Q_25Clean_Semester)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.15054 0.11653 18.455 <2e-16 ***
## FallSpringSpring 0.04415 0.15733 0.281 0.779
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.262791)
##
## Null deviance: 257.71 on 205 degrees of freedom
## Residual deviance: 257.61 on 204 degrees of freedom
## AIC: 636.66
##
## Number of Fisher Scoring iterations: 2Q10TextClean$Question[3] # p = 0.000878, Estimate = 0.4662## [1] "A lab or project where no one knows the outcome"Q10TextClean$Question[4] # p = 0.106, Estimate = 0.17299## [1] "At least one project that is assigned and structured by the instructor"Q10TextClean$Question[6] # p = 0.107, Estimate = 0.2634## [1] "A project entirely of student design"Q10TextClean$Question[11] # p = 3.37e-08, Estimate = 0.70836## [1] "Read primary scientific literature"Q10TextClean$Question[12] # p = 0.0146, Estimate = 0.3751## [1] "Write a research proposal"Q10TextClean$Question[18] # p = 0.00017, Estimate = 0.5295## [1] "Critique the work of other students"Q10TextClean$Question[19] # p = 0.0158, Estimate = 0.3499## [1] "Listen to lectures"After discussion with Moria, we would like to show, as a supplemental table the questions that have p.adjust < 0.05.
Table S4
stats_table <- tibble(Question = Q10TextClean$Question[11],
Estimate = Q_11_pre_model$coefficients[[2]],
p = coef(summary(Q_11_pre_model))[,4][[2]],
`p.adj` = p.adjust(coef(summary(Q_11_pre_model))[,4][[2]], n = 25))
stats_table <- stats_table %>%
add_row(Question = Q10TextClean$Question[18],
Estimate = Q_18_pre_model$coefficients[[2]],
p = coef(summary(Q_18_pre_model))[,4][[2]],
`p.adj` = p.adjust(coef(summary(Q_18_pre_model))[,4][[2]], n = 25))
stats_table <- stats_table %>%
add_row(Question = Q10TextClean$Question[3],
Estimate = Q_3_pre_model$coefficients[[2]],
p = coef(summary(Q_3_pre_model))[,4][[2]],
`p.adj` = p.adjust(coef(summary(Q_3_pre_model))[,4][[2]], n = 25))
kable(stats_table, "simple")| Question | Estimate | p | p.adj |
|---|---|---|---|
| Read primary scientific literature | 0.7083607 | 0.0000000 | 0.0000008 |
| Critique the work of other students | 0.5294751 | 0.0001699 | 0.0042470 |
| A lab or project where no one knows the outcome | 0.4662466 | 0.0008776 | 0.0219398 |